home *** CD-ROM | disk | FTP | other *** search
Text File | 1988-05-03 | 132.7 KB | 5,529 lines |
-
-
-
-
-
-
-
-
- User Manual for Queue Modeling Tool (Tool 37)
-
-
-
- Prepared for:
-
- NAVAL OCEAN SYSTEMS CENTER (NOSC)
- United States Navy
- San Diego, CA 92152
-
- Contract number: N66001-85-C-0040
-
- Date: June 11, 1985
-
- Ford Aerospace and Communications Corp.
- Western Development Laboratories Division
- 3939 Fabian Way
- Palo Alto, CA 94303
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 1. Introduction
-
- This user's manual explains the operation of the queueing
- modeling tool, including motivation for using the tool, a
- description of the closed models handled by this tool, and an
- explanation of the inputs and outputs produced by this tool.
- There is an on-line "help" command available with this tool.
- After invoking the tool from the operating system level, the user
- can type "help" to see all the commands available, and can type
- "help command_name" (for a specific command_name) to see details
- of any command. The user also has the ability to add, modify, or
- delete information about help commands, or to add additional help
- commands, or delete help commands. This is explained in Section
- 4 below, in the subsection entitled "Modifying Help Files".
-
- The Queue Modeling Tool is intended to be a valuable
- analytical tool for predicting bottlenecking and other
- statistical information for small models. The tool is limited to
- models where the bottlenecking is not too severe. For large
- bottlenecks, the underlying equations may produce values that are
- beyond the floating point capacity of the target system.
- Depending on the nature of the problem, the tool will do one of
- the following:
-
- 1. May abort further calculations, notify the user to reduce
- the bottleneck, and return to the MMI to allow the user
- to adjust the model.
-
- 2. May try an alternative algorithm.
-
- 3. May notify user that an UNSTABLE_SOLUTION has been
- reached, try a simple correction, and then continue. The
- user must then verify that this fix has worked by
- examining several of the reports(see 'help fix_ql').
-
- It is recommended that the modeler first begin running his
- model with a small number of jobs. The modeler can then
- experimently increase the number of jobs until the floating point
- capacity is reached. The maximun allowable number of jobs is a
- compiler parameter and is currently set to 100 jobs.
-
- Since the floating point capacity of the portable math
- library is smaller than the Data General's version, it is
- recommended that when this tool is installed on a Data General
- system, that the DG's math library, or equivalent, be used(see
- 4.3.2 Installing Machine Dependent Library).
-
- Section 4 contains details about installing this tool.
- Section 5 explains some of the state of the art uses of Ada that
- were utilized in this tool. Section 6 contains a description of
- the closed queueing model, including the parameters that the user
-
-
- Item 0037 Ford Aerospace &
- June 1985 1-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- enters to describe a given model. Section 7 explains the output
- reports produced by this tool. All the reports are available
- after the user has "run" the model, but are displayed to either a
- screen or a file only if the user issues the "report" command.
-
- Appendix A provides the help files, ie, the information that
- the user will see at the screen in response to a "help cmd" for
- any given cmd. Appendix B contains the queueing equations that
- form the mathematical basis of this model. The equations have
- been formatted for a laser printer (using troff at the
- contractor's development machine), and may not be easily readable
- in electronic form. Appendix C contains test cases, including
- input and output, so that the user may verify that the installed
- version of this tool is working correctly.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 1-002 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 2. Computer Program System Capabilities
-
- 2.1 Purpose
-
- The purpose of this interactive tool is to enable the
- program manager to model analytically (ie, mathematically, as
- opposed to discrete event simulation) the capacities of a given
- set of resources to handle a given set of jobs (which require
- services from the resources). This tool computes numerous
- steady-state statistics showing utilization of the resources,
- queue lengths, mean service times, paths taken by a job in moving
- from resource to resource, etc. An explanation of the types of
- systems amenable to modeling by this tool, as well as a detailed
- list of output reports is given below. These statistics are
- useful to determine bottlenecks in a system, to predict the
- effect of adding new resources, to predict the effect of adding
- more jobs to a given set of resources, to predict the effect of
- reconfiguring the system, etc.
-
- 2.2 General Description
-
- This tool consists of two main parts: the MMI and the
- computational part. After the user has invoked this tool, he is
- interacting with the MMI. The MMI allows the user to enter the
- description of a given system (either from the terminal directly,
- or from a file, or by a combination of these two methods), to
- save the description of a given system, to run the computations
- against a given system, and to produce reports showing the
- results of the computation.
-
- There is a comprehensive "help" command allowing the user to
- learn about which commands are available, what they do, and how
- to invoke them. The help command is virtually an on-line user's
- manual. The user may abbreviate any command or portion of
- command to the minimal extent possible. If there is any
- ambiguity the user will receive a message to that effect, and may
- reenter the command with one or more additional characters to
- resolve the ambiguity. The user may mingle upper and lower case
- characters freely without any impact on this tool. The user is
- given freedom in entering numerical values without worrying if
- they are integer or float. That is, a floating point zero may be
- entered either as 0 or as 0.0. A value of 0.3 may be entered
- simply as .3. In entering a set of values on a command line, the
- user may separate the values by blanks, commas, equal signs,
- newlines, or any combination. The Rolm AdaTM compiler on the ADE
- does not provide a backspace capability to the user entering
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 2-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- information from a terminal. We consider this a serious omission
- from the compiler. This tool provides the equivalent of a
- backspace. The user simply types the pound sign "#" to simulate
- a backspace.
-
- 2.3 Functions Performed
-
- The functions mentioned below are performed; all of them
- are invoked by issuing the name of the function from the "edit"
- mode of the MMI. Detailed descriptions of each function are
- available via the on-line help command, and are also given below
- in this user's manual.
-
- Before providing a list of the functions, it will be useful
- to explain the various modes of the MMI. The three modes are
- "edit", "prompt", and "file". The edit mode allows the user to
- issue any command in any order. If the command does not make
- sense (eg, if another command should have been issued first) the
- user will be notified. The user may re-enter a statement that
- was previously entered incorrectly, or re-enter it simply to
- change the value of some parameter in the model. The edit
- command also allows the user to go directly to either of the
- other two modes, the prompt mode or the file mode. In prompt
- mode, the tool prompts the user for a description of the system
- to be modelled. The user must enter a description of the entire
- system before leaving prompt mode. From the prompt mode, there
- is no direct way for the user to get to either of the other two
- modes. This tool knows when the user has completed describing
- the system in prompt mode, since the user must enter the number
- of nodes in the system near the beginning of the prompt session.
- Once the user has described every node in the system, he is
- automatically returned to edit mode.
-
- The third mode of the MMI is file mode. This allows the
- user to read in the description of a system from a file. At the
- completion of reading the file, the user is automatically
- returned to edit mode. The file is created by issuing the "save"
- command.
-
- The user can determine which mode he is in by the system
- level prompt. For edit mode the prompt is "E>"; for prompt mode
- the system prompt is "P>". For file mode the prompt is "I>".
-
-
-
-
- __________
-
- TM Ada is a registered trademark of the U.S.
- Government, Ada Joint Program Office
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 2-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- One implication of using these modes is that if the user is
- about to enter the description of a model, and he wishes to use
- prompt mode, he should be sure that there is enough time to enter
- the entire description during that terminal session. Otherwise
- the partial description he has entered will be lost. If there is
- a possibility that the user will not have enough time during that
- terminal session to enter all the required information, the user
- should use edit mode, since the partial results entered may be
- saved to a file at any time.
-
- The following is a list of the commands available, and a
- brief description of them.
-
- A. Help - displays on-line information about all commands.
-
- B. Run - invoke the computational part of the tool applied to
- the "current model".
-
- C. Prompt - change the MMI to prompt mode, which will then
- prompt the user for the entire description of a model.
-
- D. Njobs - specify the number of jobs in the model.
-
- E. Nnodes - specify the number of nodes in the system.
-
- F. Order - enter the names of all the nodes in the system in
- the desired order.
-
- G. Node - enter information about a particular node.
-
- H. Discipline - specify the service discipline at a given
- node. The four allowable disciplines are: first come,
- first served; processor sharing; preemptive last come,
- first served; and no-queueing. Details about these
- disciplines are given below.
-
- I. Distribution - specify the distribution function for
- providing service at each node. The three distributions
- are: exponential; R-stage erlang; R-stage coxian.
-
- J. Nservers - specify the number of servers at a given node.
-
- K. Endnode - tell the tool that you have completed specifying
- information about a given node. The tool will prompt you
- if further information is needed about the node, and will
- not execute endnode until all such information has been
- provided.
-
- L. Pbranch - specify the probability of a job moving from a
- given node to any of the other nodes in the system.
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 2-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- M. Nstages - specify the number of service stages at a given
- node. Applicable only if the distribution is erlang or
- coxian.
-
- N. Rates - specify the service rate at a given node. For a
- coxian distribution, there will be a rate for each service
- stage at that node.
-
- O. Cbranch - For a coxian distribution at a given node,
- specify the probability of branching to the next stage.
-
- P. Save - save the description of the model to a file.
-
- Q. Show - display to the terminal information about Njobs, a
- given node, the entire model, Nnodes, Order, or Title.
-
- R. Quit - leave this tool.
-
- S. Infile - read in the (full or partial) description of a
- system from a file.
-
- T. Outfile - specify to which file reports are to go.
-
- U. Reset - Wipe out the description of the current model.
-
- V. Report - Generate a specified report based on having run
- the computations against a given model.
-
- W. Title - Specify a title, which will be output as part of
- the header information on each report generated.
-
- X. Echo - specify whether reports that are generated should be
- displayed at the terminal.
-
- Y. Paging - specify whether header information should be
- displayed to the terminal for every page of the report.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 2-004 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 3. Notification of Future Enhancements
-
- From time to time Ford Aerospace will enhance this queueing
- modeling tool. User's wishing to receive notification of such
- enhancements should send their name and address to:
-
- Division Software Technology and Support
- Ford Aerospace and Communications Corp.
- Western Development Laboratories Division
- 3939 Fabian Way
- Palo Alto, CA 94303
- ATTN: Ada Tools Group
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 3-001 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 4. Installing Queue Modeling Tool
-
- 4.1 Deliverable Items
-
- The following files comprise the deliverable items for the
- tool.
-
- qsap_1.ada -- 3 source files with sources arranged
- qsap_2.ada -- in an acceptable compilation order.
- qsap.ada -- main program
-
- help.ada -- the path name pointing to the help
- -- directory must be changed for each
- -- installation. See Section below
- -- "Editing and Recompiling package body
- -- Help_Setup".
-
- xxxxx.hlp -- Approximately 30 files are being delivered
- -- as part of this contract with names
- -- of commands replacing "xxxxx", eg,
- -- order.hlp, save.hlp, etc.
-
- math_dg.ada -- package body of Gen_Math which uses
- -- Data General's math lib.
-
- Test_1.dat -- Test files
- Test_2.dat
-
-
-
- 4.2 Compiling, Linking, and Executing the Tool
-
- The sources comprising the tool have been concatenated into
- three files. These modules were arranged in an acceptable
- compilation order and do not use any machine dependent features.
- This tool was compiled, linked, and executed in a Data General
- environment using the following commands.
-
- > ada qsap_1.ada -- Compile qsap sources.
- > ada qsap_2.ada
- > ada/main_program qsap.ada
-
- > ada help.ada -- Compile application specific
- -- package body Help
-
- > adalink/root_stack_size = 250000/mtop=6 QSAP
-
- -- [Help facility should be installed
- -- prior to execution(Sec. 4.5)]
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- > x qsap -- Execute QSAP
-
-
- 4.2.1 Installation Checkout
-
- Two test_case files should be executed and the results
- compared with those in appendix C.
-
- Test_1.dat Test_2.dat
-
- > x qsap
- E> inf = test_1.dat
- -- Output is stored in file result_1.rpt
- E> quit
-
-
-
-
- Also try several of the help commands to insure that the help
- directory has been properly installed.
-
- 4.3 Math library installation
-
- 4.3.1 Machine Independent version of the Tool
-
- The tool intensively uses the generic math package Gen_Math
- which contains SQRT, EXP, LOG, functions. Since there are no
- standardized Ada math libraries, a portable public domain math
- library (written in Ada) was obtained and imported to the body of
- package Gen_Math. The portable packages are:
-
- package Floating_Characteristics
- package Numeric_IO
- package Numeric_Primitives
- package Core_Functions
-
-
- Since the portable math library was written in Ada without
- knowledge of the underlying hardware, these routines run
- significantly slower and less accurate than system dependent math
- libraries. However for small models without severe bottlenecks
- this portable library may be entirely satisfactory. The file
- qsap_1.ada contains the above packages and hence no additional
- installation effort (beyond 4.1) is required.
-
- 4.3.2 Machine Dependent Math Library
-
- For very large models, a more efficient and accurate math
- library may be desirable. The most direct approach is to modify
- the package body of Gen_Math so that it uses a target system math
- library. The file Gen_Math_Body_DG.ada illustrates how the Data
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- General ADE math library was attached. To incorporate this into
- the tool, one must replace the portable version (in file
- qsap_tool.ada) with this special version and remove the packages
- listed in section 4.3.1.
-
- --========== math_dg.ada =======
- with Math_Lib; use Math_Lib;
- package body Gen_Math is
- --============================================================
- -- Imports Data General's Math Lib(Assembly Routines).
- --
- -- Installation Notes:
- -- o Use ADE LibsearchList command to point to
- -- DG's package Math_Lib.
- -- o Must replace portable package body Gen_Math in
- -- file qsap_tool.ada with this version
- -- and recompile system ( since Gen_Math
- -- is a generic package).
- --
- --=============================================================
-
- --*******************************************
- function Log( X : FltType) return FltType is
- begin
- return FltType(Logarithm(Float(X)));
- end Log;
-
- --*******************************************
- function Sqrt( X : FltType ) return FltType is
- begin
- return FltType(Square_Root(Float(X)));
- end Sqrt;
-
- --*******************************************
- function Exp ( X : FltType ) return FltType is
- begin
- return FltType(Exponential(Float(X)));
- end Exp;
- end Gen_Math;
-
- 4.4 Linking the Tool
-
- This tool extensively uses matrices in which the maximum
- alloted index size is a program parameter. A change to this
- parameter MAX_INDEX_SIZE (in package Network_Parameters) will
- significantly effect the size of memory required for the
- executable module. Shown below is the ADE link command when
- MAX_INDEX_SIZE is set to 100 .
-
- Adalink/root_stack_size = 250000/mtop=6 QSAP
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 4.5 Installing The Help Facility
-
- This tool's help facility is split into numerous help files
- which must reside in a single directory. If the location of the
- help directory is to be different than the directory of the tool,
- then the package body of Help_Setup must be edited to specify
- this help directory.
-
- 4.5.1 Installing the Help directory and Help files
-
- Installing the Help directory, consists of making a
- directory and copying the help files into this directory. For
- delivery all help files have names ending with ".hlp".
-
- 4.5.2 Editing and Recompiling 'package body Help_Setup'
-
- To allow this tool to access the help directory, the source
- of package body Help_Setup must be modified (as shown below). To
- facilitate this change, a special Ada file (Help.ada) is
- supplied. Change the path name (:udd:facc_krg:testbed:help: in
- the example below) to whatever path name points to the help file
- directory in your installation.
-
- On systems where the file extension ".hlp" is inappriopriate, an
- alternate extension may be specified.
-
- If the tool has been previously compiled according to
- section 4.2, then only this body needs to be recompiled and a new
- link module created.
-
- > Ada help.ada
- > Adalink/root_stack_size = 250000/mtop=6 QSAP
-
- --=========File: Help.ada
-
- package body Help_Setup is
- begin
- --===============Help_File_Extension
- Help_File_Name_Ext := Txt( ".hlp");
-
- --=======================Help Directory ============
- Help_Directory := Txt(":udd:facc_krg:testbed:help:");
-
- end Help_Setup;
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-004 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 4.5.3 Deliverable Help files.
-
- See Appendix A.
-
- 4.5.4 Modifying Help Files.
-
- One can to modify, add, and delete help files. The index
- file index.hlp must be manually updated whenever a help file is
- added or deleted. Note: the help names in file index.hlp must
- match with the external help file names except for the filename
- extension (.hlp). For examle, to add a new help file, new.hlp,
- the user first creates this file using any desired text editor,
- in the appropriate help directory. Next, place the file name
- "new" in alphabetical order in the file index.hlp (in the command
- sequence following the word "%INDEX"). Then when the user types
- "help new" during execution of this tool, the contents of file
- new.hlp will be displayed to the screen.
-
- Likely candidates for new help files are user-specific
- information, and bug reports, etc.
-
- 4.6 Exceptions
-
- During execution all uncorrectable exceptions are reported
- and control is returned to the MMI module. The user may then wish
- to quit or attempt to correct the situation. Several classes of
- exceptions are listed below:
-
- MATRIX_INVERSION_ERROR:
- The matrix of branching probabilities can not be inverted.
-
- Corrective Action: Change branching probs.
-
- UNSTABLE_SOLUTION_ERROR:
- Occurs whenever the result of a calculation is an
- unrealistic value (.e.g. Prob <0.0 or >1.0 ). The equations
- described in appendix B may yield values that exceed
- the accuracy or capacity of the implementation. Depending on
- the equation, this tool may attempt an alternate equation, may
- notify user that a value is erronous, try to
- correct value, and continue processing, or
- give up, raise the Unstable_Solution_Error exception, and return
- back to the MMI.
-
- Corrective Action: Try changing input values such as service
- rates and num of jobs (njobs). Increase the former and
- decrease the latter. Severe bottlenecks are likely candidates
- for this type of error.
-
- Unknown Exception:
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-005 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Any other exception will be reported as an unknown exception.
-
- Corrective Action: Hopefully, the user will be able to retrace
- his steps and find the error. If the installation is
- maintaining a bug help file, try using the
- help command.
-
-
- 4.7 Compilation Units
-
- The compilation units comprising the this tool are shown in
- the following table.
-
- Compilation Units
- =======================
-
- -- Setup --
-
- package Help_Setup ! Help Facility Installation.
- package Network_Parameters ! Network and Matrix size parms.
-
- procedure Qsap ! Main Program
-
- -- MMI Subsystem --
-
- package MMI ! User Interface.
- package MMI_IO ! IO facility.
-
- -- Global Data Access Subsystem --
-
- package Global_Types ! System Global Types
- package Network ! Data Store for Network of Nodes.
- package Node_Servicer ! Abstract type NodeDef
-
- -- Statistical Subsystem
-
- package Net_Stats ! Statistical Routines Specification
- package Net_Data_Pak ! Local Data Storage for Net_Stats.
-
- -- Report Generation Subsystem
-
- package Report_Types ! Types for Reports & Net_Stats.
- package Report_Lists ! Linked Lists for Reports & Net_Stats.
- package Reports ! Report Routines
-
- -- Library Packages --
-
- package Gen_Text_Handler ! Generic Text Handler( Abstract type)
- package Gen_List_Handler ! Generic List Handler( Abstract type)
- package Gen_Dyn_Mat ! Generic Dynamic Matrix(Abstract type)
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-006 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- package Gen_Math ! Generic Math routines(Pack Spec only)
- package Gen_Factorials ! Table_access Generic Factorial Routines.
- package Min_Max_Pak ! Min Max Routines.
-
- package Float_Pak ! Machine Indep. Float Type
- package Core_Functions ! Machine Indep. Math Routines.
- package body Gen_Math_Body ! Machine Indep. body of Gen_Math.
- package body Gen_Math_Body ! DG version of body of Gen_Math.
-
- -- Tool Specific Library Packages
-
- package Real_Mat_Pak ! Instantiation of Gen_Dyn_Mat.
- package Text_Handler ! Instantiation of Gen_Text_Handler.
-
- 4.8 Compilation Order
-
- The sources in file qsap_tool.ada have been arranged to
- produced following compilation order. These units have also been
- arranged so that subunits immediately follow their parent unit(
- indentation shows subunit structure). The following table shows
- this compilation order.
-
- Tool Compilation Order
- ----------------------
-
-
- Package NETWORK_PARAMETERS
-
- Package MIN_MAX_PAK
- Package body MIN_MAX_PAK
-
- Generic package GEN_TEXT_HANDLER
- Package body GEN_TEXT_HANDLER
-
- Generic package GEN_LIST_HANDLER
- Package body GEN_LIST_HANDLER
-
- Generic package GEN_DYN_MAT
- Package body GEN_DYN_MAT
-
- Package GLOBAL_TYPES
- Package body GLOBAL_TYPES
-
- Package TEXT_HANDLER
-
- Package REAL_MAT_PAK
- Package body REAL_MAT_PAK
-
- Package FLOATING_CHARACTERISTICS
- Package body FLOATING_CHARACTERISTICS
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-007 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Package NUMERIC_IO
- Package body NUMERIC_IO
-
- Package NUMERIC_PRIMITIVES
- Package body NUMERIC_PRIMITIVES
-
- Package CORE_FUNCTIONS
- Package body CORE_FUNCTIONS
-
- Generic package GEN_MATH
- Package body GEN_MATH
-
- Generic package GEN_FACTORIALS
- Package body GEN_FACTORIALS
-
- Package NODE_SERVICER
- Package body NODE_SERVICER
-
- Package NETWORK
- Package body NETWORK
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-008 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Package MMI_IO
- Package body MMI_IO
-
- Package REPORT_TYPES
-
- Package REPORT_LISTS
-
- Package REPORTS
- Package body REPORTS
-
- Package NET_STATS
-
- Package NET_DATA_PAK
- Package body NET_DATA_PAK
-
- Package body NET_STATS
- Function body .IS_MATCH
- Function body .MAKE_INDEX_LIST
- Function body .QUEUE_SIGMA
- Function body .FCFS_SIGMA1
- Function body .FCFS_SIGMA2
- Function body .CALCULATE_FCFS_PHI
- Procedure body .DISPLAY_PBRANCH
- Procedure body .DISPLAY_ARRIVAL_FREQS
- Function body ..SUM_ALPHAS
- Procedure body .DISPLAY_SERV_TIMES
- Procedure body .DISPLAY_Q_LENGTHS
- Procedure body .DISPLAY_RESPONSE_TIMES
- Procedure body .DISPLAY_GNORMS
- Procedure body .DISPLAY_Q_LENGTH_DISTS
- Procedure body .DISPLAY_ROUTING
- Procedure body .DISPLAY_SERVICE
-
- Procedure body .CALCULATE_STATS
- Procedure body ..INITIALIZE_PB_MAT
- Procedure body ..INITIALIZE_NODE_ARRAY
- Procedure body ..CALCULATE_NETWORK_MOMENTS
- Function body ...EXP_MOMENTS
- Function body ...ERLANG_MOMENTS
- Function body ...COXIAN_MOMENTS
- Function body ....MAKE_V_DELTA
- Function body ....COXIAN_SIGMA
- Procedure body ..CALCULATE_STEADY_STATE_FLOW
- Function body ...BUILD_P_STAR
- Procedure body ..CALCULATE_EQUILB
- Function body ...SIGMA
- Function body ...SELECT_MUS
- Function body ...SELECT_ALPHAS
- Procedure body ..CALCULATE_GNORMS
- Procedure body ..CALCULATE_AUX_ARRAY
- Procedure body ..CALCULATE_Q_LENGTH_DIST
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-009 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Procedure body ..CALCULATE_THRU_PUT
- Procedure body ..CALCULATE_UTILIZATION
- Procedure body ..CALCULATE_MEAN_RESPONSE
-
- Package HELP_SETUP
- Package body HELP_SETUP
-
- Package MMI
- Package body MMI
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 4-010 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 5. State of Art Uses of Ada
-
- 5.1 Ada Features
-
- Emphasis was placed on developing highly readible and
- structured Ada code. Almost all of Ada features were use (except
- for tasking and machine dependent features). Listed below are
- some of interesting Ada features and examples of their usage.
-
- Enumeration Types: package MMI,
- package Global_Types
-
- Access Types: package Network,
- package Gen_List_Handler
-
- Private Types: package Gen_Text_Handler
- package Gen_List_Handler
- package Gen_Dyn_Mat
- package Node_Servicer
-
- Discriminant Records package MMI
- package Global_Types
- package Gen_Dyn_Mat
- package Gen_Text_Handler
- package Net_Data_Pak
-
- Variant Records package Global_Types
- package Report_Types
-
- Generics package Gen_List_Handler
- package Gen_Text_Handler
- package Gen_Dyn_Mat
- package MMI
-
- Exceptions package MMI
- package Net_Stats
- package Calculate_Stats
- package Gen_Dyn_Mat
-
- Stubs and Subunits package body Net_Stats
- procedure body Calculate_Stats
-
-
- 5.2 Naming Conventions
-
- To reduce the naming complexity, the following naming
- conventions were used.
-
-
- o reserved words; lower case
-
-
- Item 0037 Ford Aerospace &
- June 1985 5-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- o Types; First and each significant character were capitalized.
- Underscores separated meaningful words, eg:
-
- End_Of_Network : Boolean;
- procedure Calculate_Thru_Put;
- Mat_Init_Exception : exception;
-
- o Objects; First and each significant character were capitalized.
- No Underscores separated meaningful words, eg:
-
- type NumServers is ...
- type NodeDef is ....
- type RealMatrix is ...
- Num_Servers : NumServers;
-
- Three advantages to the above notation were:
- (1) types easily distinguishable from objects,
- (2) type names easily deduced from object names, and
- (3) reduction of the number of unique identifiers.
-
-
- 5.3 Pretty_Printing
-
- Considerable effort was placed in developing highly readible
- code. Automatic Pretty_printing should be avoided since it will
- reduce program readibility for the following reasons.
-
- (1) type names will become less readible if case conventions
- are changed ( NumNodes becomes NUMNODES).
-
- (2) Page breaks have been carefully inserted to improve
- readibility.
-
- (3) Equations and comments have been carefully alligned to
- to improve readibility.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 5-002 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 6. Description of the Closed Queueing Model
-
- This tool allows any system to be modeled, in which the
- system is expressible as a fixed number of "jobs" circulating
- among a fixed number of "nodes" (or servers). Such a system is
- called a "closed model" if the number of jobs is specified in
- advance, and new jobs may neither enter nor leave the system.
- The basic purpose of this tool is to compute a variety of
- steady-state statistics for this closed model. For example, if
- there are 20 computer terminals hooked to a main frame computer
- and network server, the 20 terminals might represent jobs, and
- the computer and network server might represent nodes. The
- following are the characteristics of interest of such a system,
- and represent the user-supplied information:
-
- A. Number of jobs.
-
- B. Number of nodes.
-
- C. Branching probabilities, ie, the probabilities that if a
- given job is currently at node i, that job will next
- require service at node j.
-
- D. Description of each node:
-
- i. The "service discipline" at the given node. There
- are four types of service that can be modeled, any
- one of which is available at a given node. Once a
- service discipline is specified for a node, it
- remains the same during the computational part of a
- given session. The user may change the service
- discipline at a given node prior to invoking the
- computational module. The four service disciplines
- are as follows.
-
- A. First come, first served (FCFS)
-
- B. Processor sharing: Any job which arrives at
- the node gets instantaneous service, but at a
- rate proportional to the number of jobs
- present. In particular, the rate at which any
- job receives service is 1/K times the service
- rate of the processor, where K is the number of
- jobs currently at the node.
-
- C. Preemptive resume, last come, first served:
- the lastest arriving job gets immediate and top
- priority. This is a variation of the
- traditional LCFS, or stack discipline.
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 6-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- D. No queueing: This discipline assumes that
- there is enough capacity at the given node to
- provide instantaneous service to any job
- entering that node, without any diminution in
- service rate provided to other jobs already at
- that node.
-
- ii. The distribution at each node. This random variable
- represents how much service a job will require at the
- given node. The three distributions handled by the
- model are as follows:
-
- A. Exponential: this traditional distribution
- usually represents a worst-case assumption and
- often leads to great simplification in terms of
- representing the user's problem space. It
- should be assumed when better information is
- lacking.
-
- B. Erlang: utilized when the service requirements
- at each node consist of n similar stages, where
- n is fixed in advance by the user during model
- entry. For example, if the service at a node
- consists of sending a message, there might be
- three stages of service: send header, send text
- of message, send trailer. These three stages
- must all be completed for the service provided
- by the node to be deemed completed.
-
- C. Coxian: Utilized when there are multiple stages
- of service (as in Erlang), but when the stages
- are dissimilar, and when there is a chance that
- the job may be complete prior to going through
- all stages of service. The user must provide
- the number of stages, (as in Erlang), and must
- also provide the probability of transition from
- one stage to the next. Thus, if there are n
- stages, the user must provide n-1 transition
- probabilities pi. Each pi is a number greater
- than zero and less than or equal to unity. For
- example, service at a given node might consist
- of the sequence compile, link, run. At the
- early stages of a project, the probability of
- going from the compile to the link stage might
- be 0.1 (ie, one out of ten chance of
- successfully compiling). At a later stage of
- the project, this probability might rise to
- 0.9.
-
- In summary, for the Coxian distribution,
- the user must enter the number of stages, the
-
-
- Item 0037 Ford Aerospace &
- June 1985 6-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- service rate parameter for each stage, and the
- branching probabilities between stages.
-
- iii. The service rate: This is a constant provided by the
- user stating how much service can be provided by a
- given node per unit time. For exponential and erlang
- distributions, there is a single service rate entered
- for the node. In the erlang case, this single rate
- applies to each stage. In the coxian case, the user
- provides a separate service rate for each of the
- stages of service.
-
- iv. Nservers, the number of servers at each node. For
- the first come, first served discipline, the user is
- allowed to enter the number of servers at a given
- node. All servers provide service at the same user-
- specified rate. For processor sharing, and for the
- preemptive last come, first served disciplines, the
- number of servers at the node is automatically set to
- 1 by this tool; thus the user does not enter a value
- for Nservers. For the no queueing discipline, the
- number of servers is automatically set to the number
- of jobs (Njobs) previously entered by the user.
- Thus, the user does not enter a value for Nservers
- for the no queueing discipline.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 6-003 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 7. Output Reports Produced
-
- Ten reports are available after the computational part of
- this tool has been applied to a model description. These reports
- are as follows, and are described below:
-
- 1. Arrival_frequencies
-
- 2. Serv_Times
-
- 3. Response_Times
-
- 4. Qlength_Distributions
-
- 5. Throughput
-
- 6. Routing
-
- 7. Serv_Requirements
-
- 8. Model
-
- 9. Normalizations
-
- 10. Pbranch
-
- The user has the option of specifying which of the reports
- are to be produced, and whether the reports are to be written to
- a file, displayed at the terminal screen, or both. If the user
- does not specify a given report, the information that would have
- been available is lost, ie, there are no files that are produced
- automatically. Of course, the user can always rerun the job if
- this happens.
-
- All the reports contain header information including the
- date the report was produced. It is recommended that the user
- produce all reports of interest from a given run to a single
- file, and that the first report on this file be the description
- of the model inputted by the user, so that there will be a clear
- correspondence between inputs and outputs on any given file.
-
- 7.1 Model Inputted by the User
-
- There are four reports that allow the user to see the model
- that has been input. These are:
-
- 1. Model: Shows the definition of the model on which the
- computational part was run. (To produce this report type
- "report model".)
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 7-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 2. Branch probabilities: When the user entered the model, he
- provided branch probabilites from a given node to all other
- nodes. Therefore, the user entered the branch in a
- "scattered" fashion, embedded within the various node
- definitions. This report provides a single, centralized
- matrix of these branch probabilities. (To produce this
- report type "report pbranch".)
-
- 3. Relative Arrival Frequencies: Provides a summary of the
- node names and the queue discipline entered by the user,
- plus the relative arrival frequencies at each node as
- computed by the tool. (To produce this report type "report
- arrival_frequencies".)
-
- 4. Service times: Provides the service distribution for each
- node, as entered by the user, plus the mean service time,
- service time variance, and coefficient of variation. The
- service time is the inverse of the service rate, which had
- been entered by the user. (To produce this report type
- "report serv_times".)
-
- 7.2 Queue Lengths
-
- This provides the following types of information at each
- node.
-
- 1. Mean queue length: This statistic includes jobs being
- served by the node, as well as jobs waiting for service.
- The sum of mean queue lengths over all the nodes equals the
- number of jobs in the system.
-
- 2. The variance and coefficient of variation of the mean queue
- length.
-
- 3. Throughput: The mean number of jobs processed by the node
- per unit time.
-
- 4. Utilization: The fraction of time the node is busy.
- (To produce this report type "report throughput".)
-
- 7.3 Response Times
-
- This report provides the following types of information at
- each node.
-
- 1. Mean response time: The waiting plus service time
- experienced by an arbitrary job at the node.
-
- 2. The variance and coefficient of variation of the response
- time.
- (To produce this report type "report response_times".)
-
-
- Item 0037 Ford Aerospace &
- June 1985 7-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 7.4 Normalization Constants
-
- This report provides normalization constants at each node.
- These are important internal variables used in computing
- statistics in the other reports, and not generally of interest to
- the end user. (To produce this report type "report
- normalizations".)
-
- 7.5 Queue Length Distributions
-
- Shows for each node the probability that exactly n jobs will
- be at that node at a random point in time, for each value of n
- from 1 to number of jobs. The number of lines of output in this
- report is therefore on the order of (number of nodes) * (number
- of jobs). (To produce this report type "report
- qlength_distributions".)
-
- 7.6 Routing Behavior
-
- As a job branches from node i to node j, the user might be
- interested in knowing how many times a given node k was visited
- in the interim, where k /= j. This report provides this
- information (as well as the variance) for any set of (i,j,k) in
- which the user is interested. Since in the extreme case the
- number of lines of output is on the order of the cube of the
- number of nodes, the user is given the option of restricting the
- amount of output by restricting the "from", "to", and "by" sets
- of nodes. (To produce this report type "report routing".)
-
- 7.7 Node Service Requirements and Return Times
-
- A job arrives at a given node j, receives service at node j,
- visits other nodes, and then arrives at a specified node k.
- This report produces the following statistics for an arbitrary
- job along this path, for all desired nodes other than the
- destination node k.
-
- A. Service Requirement Mean: this includes only service, not
- waiting times.
-
- B. Service Requirement Variance.
-
- C. Residence Time Mean: this includes service plus waiting
- time at the given node.
-
- D. Total Service Requirement Mean (before a job returns to
- node k, given that the job starts at node j): this includes
- the total amount of service (not counting waiting) that the
- job encountered along the path from node j to node k.
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 7-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- E. Total Residence Time Mean: This is the average time
- required for a job to return to node k, given that the job
- starts from node j: this includes the total amount of
- service plus waiting time that the job encountered along
- the path from node j to node k.
- (To produce this report type "report serv_requirements".)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 7-004 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 8. Appendix A - Description of Commands
-
- This appendix contains the same information that is
- available to the user via the on-line help command. By typing
- "help", the user will see all the commands available. Then the
- user can type "help command_name" to get information about any
- given command. Details of the various commands follow.
-
- This help file must be modified by the following rules.
-
- * All text before '%INDEX' is treated as comments.
-
- * The first occurance of '%' on any line in column 1 must be
- the start of the index.
-
- * The index consists of '%INDEX' starting in column 1 and the
- help topics following one per line and also starting in
- column 1. The index is terminated by a '%END' in column 1
- of a subsequent line.
-
- * A help topic must be listed in the index.
-
- * A help topic will be located in a file with the same name.
- In this file a help topic named 'TOPIC' will start with
- '%TOPIC' in column 1 and end with either '%' encountered in
- column 1 on a subsequent line or end of file.
-
- * Topic names are limited in length. For the index to print
- properly, a name should not be more than 19 characters.
- However, if the system cannot accomodate 19 characters in a
- file name, then they will have to be shorter.
-
- * Topic text is displayed all at once, so the topics should
- be small enough to fit on one page.
-
- %INDEX
- Batch
- Cbranch
- DISCipl
- DISTrib
- ECho
- ENdnode
- EXAmple
- EXCepts
- Fix_ql
- Help
- Infile
- NEw_User
- NOde
- NNodes
- NJobs
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NSErvers
- NSTages
- ORder
- OUtfile
- PAging
- PBranch
- PRompt
- Quit
- RAtes
- REPort
- RESet
- RUn
- SAve
- Show
- Title
- %END
-
- BATCH
-
- To run the program in batch mode, Choose one of the
- following examples pertaining to the system you are running on.
-
- DATA GENERAL:
-
- -) csh -- get into the Unix C-Shell
- % qbatch -i -s /notify -- submit a batch job where INPUT
- x qsap < INPUT > OUTPUT -- is the input file name and
- ^d -- OUTPUT is the output file name
- % exit -- where the results are to go.
- %END
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- CBRANCH
-
- SYNTAX
-
- Cbranch [=] {probability_value}
-
- DESCRIPTION
-
- Cbranch is used to specify the continuation probabilities between
- stages for a node with an R-Stage Coxian service distribution. The
- number of probability values must be one less than the number of
- stages and each value must be in the range 0.0 .. 1.0 excluding 0.0.
-
- EXAMPLES
-
- node = abc node def
- dist = coxian dist coxian
- nstages = 4 nstages 2
- cbranch = .5, 0.3 .2 c .75 -- minimum abbreviation
-
- PREREQUISITES
-
- Nnodes, order, njobs, node, distribution = coxian, nstages.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- DISCIPLINE
-
- SYNTAX
-
- DISCipline [=] (Fcfs | Nq | PR_lcfs | P_share)
-
- DESCRIPTION
-
- Discipline is used to specify the service discipline at a node. If
- a fcfs discipline is chosen then the distribution is automatically
- set to exponential. If a nq discipline is chosen, the number of
- servers is set equal to the number of jobs. If a p_share or pr_lcfs
- discipline is chosen, the number of servers is set to 1.
-
- EXAMPLES
-
- node = a
- discipline = fcfs -- sets distribution to exponential
- disc p_ -- resets discip and sets nservers = 1
-
- PREREQUISITES
-
- Nnodes, order, njobs, node.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-004 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- DISTRIBUTION
-
- SYNTAX
-
- DISTribution [=] (EXponential | ERlang | Coxian)
-
- DESCRIPTION
-
- Distribution is used to specify the service distribution at a node.
- For a fcfs discipline, the distribution is set to exponential.
-
- EXAMPLES
-
- node = foo node fooy
- distribution = exponential discip fcfs -- sets dist to expon
- dist erlang -- causes a warning
-
- PREREQUISITES
-
- Nnodes, order, njobs, node.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-005 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- ECHO
-
- SYNTAX
-
- ECho (Yes | No)
-
- DESCRIPTION
-
- Echo determines whether or not output reports will be displayed at
- the terminal. This decision is independent of the current outfile
- setting. The default setting is 'Echo Yes'.
-
- EXAMPLES
-
- echo no
-
- ec y -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-006 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- ENDNODE
-
- SYNTAX
-
- ENdnode
-
- DESCRIPTION
-
- Endnode is used to signal the end of a node definition block.
- Commands occurring after a 'node' command which pertain to the node
- will not be saved into the model until the endnode command is issued.
- If this command is issued when the node information is not complete,
- a message will be printed indicating which commands are missing.
-
- EXAMPLES
-
- node peter_pan node fairy
- . .
- endnode en -- minimum abbreviation
-
- PREREQUISITES
-
- Nnodes, order, njobs, node, and all node subcommands.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-007 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- EXAMPLE
-
- This contains an example model definition.
-
-
- nnodes 3 -- sets the number of nodes
- order a,b c -- nodes a,b,c in order. Comma and space are
- njobs 10 -- valid delimiters.
-
- node a -- start of node definition block
- discipline fcfs -- distribution will be set to exponential
- nservers 2
- rate 2.4 -- exponential rate
- endnode -- end of node definition block
-
- node b
- discipline p_share -- nservers will be set to 1
- dist erlang -- a unique abbreviation of distribution is ok
- nstages 3
- rate 3.45 -- this rate applies to each of the 3 stages
- end
-
- node c
- discip nq -- nservers will be set to 10 = njobs
- dist coxian
- nstages 4
- cbranch .4 .2 .3 -- nstages-1 continuation probabilities
- rates 1.4 4.5
- 2.3 5.7 -- command can extend over more than one line
- end
-
- run -- executes the model
-
- report arrival_freq all
- report serv_times (a c)
- report response_time a
- report qlength_dist (b c)
- report throughput all
- report routing from a to all by c
- report serv_requir all a c
- report normalization
- report pbranch
- report model
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-008 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- FIX_QL
-
- DESCRIPTION -- A FIX for QLENGTHS
-
- While computing equation 1.2.0.11, an erronous value for 'l' has been
- detected. To continue , this value has been set to zero. Hence,
- the computed value for qlength for this node may be suspect.
-
- CAUSE
-
- This situation can occur when there is a serious bottleneck at one
- of the nodes. The equation 1.2.0.11, in appendix B, will then
- attempt to subtract two extremely close numbers whose difference
- is beyond the accuracy of the machine. The ramifications of this
- problem may seriously distort the qlength distribution for that node.
-
- CORRECTIVE ACTION
-
- The value for 'l' has been reassigned to zero which results in
- qlength probability of zero. It appears that 'l' misbehaves
- only in the negligible reqions of the qlength distribution
- and hence assigning qlength probability to zero should
- not seriously impact any of the reports.
-
- VERIFY Q_LENGTHS
- One can easily check if the correction has worked by examining
- the Thru_Put report. If the sum of the mean queue lengths
- approximates the number of jobs, then the 'fix' worked.
- If not, then reduce the BOTTLENECK!
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-009 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- EXCEPTS
-
-
-
- DESCRIPTION -- EXCEPTIONS
-
-
- During execution any uncorrectable exception is reported
- and control is returned to the MMI module. The user may then wish
- to quit or attempt to correct the situation. Several types of
- exceptions are listed below:
-
- MATRIX_INVERSION_ERROR:
-
- The matrix of branching probabilities can not be inverted.
-
- Corrective Action: Change branching probs.
-
- UNSTABLE_SOLUTION_ERROR:
-
- Occurs whenever the result of a calculation yields an
- unrealistic value. The equations described in the appendix B
- may yield values that exceed the accuracy or capacity of the
- target system. Depending on the equation, QSAP may attempt
- an alternate solution, may report the error and continue
- processing, or give up and raise the Unstable_Solution_Error
- exception.
-
- Corrective Action: Try changing input values such as service
- rates and number ofjobs. Severe bottlenecks
- are likely candidates for this type of
- error.
-
- UNKNOWN_EXCEPTION:
-
- Any other exception will be reported as an unknown exception.
-
- Corrective Action: Hopefully, the user will be able to retrace
- his steps and find the error. If the installation is
- maintaining a bug helpfile, try the help facility.
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-010 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- HELP
-
- SYNTAX
-
- Help [help_topic]
-
- DESCRIPTION
-
- Help provides online information on commands and other features. If
- the help_topic keyword is not supplied, then a list of all help
- topics is displayed.
-
- EXAMPLES
-
- help -- gives a list of all help topics
- help nnodes -- describes the nnodes command
- h nn -- minimum abbreviation
-
- PREREQUISITES
-
- none
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-011 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- INFILE
-
- SYNTAX
-
- Infile file_name
-
- DESCRIPTION
-
- Infile is used to execute commands from a file. The file can contain
- a complete or partial model definition. The file can contain any
- command except another infile command.
-
- EXAMPLES
-
- infile = my_file -- reads from file, my_file
- i my_model -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-012 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NEW_USER
-
- This contains some important items that the user should be aware of.
-
- * The online help facility can be useful. Type 'help' for a list of
- topics and 'help topic' for information on that topic.
-
- * The command format is free. A command can span more than one line.
- More than one command can occur on a single line. The characters ',',
- '=', and space are delimiters. ORDER a,b c=d e is equivalent to
- ORDER a b c d e.
-
- * In edit mode the program will not respond to the command until all
- of the necessary command parameters have been supplied. If a command
- seems to hang it is because something is missing. The only commands
- that take more than a few seconds are 'report' and 'run'.
-
- * Abbreviated commands are possible. For example 'or a' is equavalent
- to 'order a'. The help information and the prompt mode follow the
- convention that the minimum abbreviation of a word is in upper case
- and the rest of the word is in lower case.
-
- * The character '' is a character kill (some systems cannot handle
- the backspace character). For example, caad\re is equal to care.
-
- * Real values can be entered without leading zeros or trailing decimal
- points. For example, .5 and 3 can be entered where a real is expected.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-013 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NODE
-
- SYNTAX
-
- NOde [=] node_name
-
- DESCRIPTION
-
- Node is used to signal the beginning of a node definition block. The
- node_name must match a node that was specified in the order command.
- Commands subsequent to the node command which pertain to the node
- will not be saved into the model until the endnode command is issued.
-
- EXAMPLES
-
- node capacitor no resistor -- minimum abbreviation
- . .
- . .
- endnode endnode
-
- PREREQUISITES
-
- Nnodes, order, njobs.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-014 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NNODES
-
- SYNTAX
-
- NNodes [=] natural_number
-
- DESCRIPTION
-
- Nnodes is used to specify the number of nodes in the network layout.
- It is generally the first command issued and once issued, it cannot
- be re-issued until after a reset.
-
- EXAMPLES
-
- nnodes = 10
-
- nn 2 -- minimum abbreviation
-
- PREREQUISITES
-
- Reset or initial entry into the program.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-015 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NJOBS
-
- SYNTAX
-
- NJobs [=] natural_number
-
- DESCRIPTION
-
- Njobs is used to specify the number of jobs in the network layout.
- It can be issued and re-issued at any time.
-
- EXAMPLES
-
- njobs = 13
-
- nj 14 -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-016 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NSERVERS
-
- SYNTAX
-
- NSErvers [=] natural_number
-
- DESCRIPTION
-
- Nservers is used to specify the number of servers at a node. If a
- NQ discipline is chosen then nservers is set equal to the number of
- jobs. If a PS or PR_LCFS discipline is chosen then nservers is set
- equal to 1. A FCFS discipline is the only one which does not force
- a specific value for nservers.
-
- EXAMPLES
-
- node = puff node sleepy
- discip = fcfs discip ps -- nservers will be set to 1
- nservers = 10 nse 5 -- will cause a warning
-
- PREREQUISITES
-
- Nnodes, order, njobs, node.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-017 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- NSTAGES
-
- SYNTAX
-
- NSTages [=] natural_number
-
- DESCRIPTION
-
- Nstages is used to specify the number of stages of an R-Stage Erlang
- or Coxian distribution at a node.
-
- EXAMPLES
-
- node = jiffy node zorba
- dist = erlang dist coxian
- nstages = 6 nst 3 -- minimum abbreviation
-
- PREREQUISITES
-
- Nnodes, order, njobs, node, distribution = erlang or coxian.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-018 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- ORDER
-
- SYNTAX
-
- ORder [=] [(] {node_name} [)]
-
- DESCRIPTION
-
- Order is used to specify the node names and their order in the
- network layout. A node name can be up to 15 characters in length
- and can contain any characters except ',', '=', '#' or blank. A
- node name cannot be 'all', 'from', 'by', 'to', which are reserved
- words.
-
- EXAMPLES
-
- nnodes = 4
- order (cpu, disk terminal, modem)
-
- or Bob Ted Carol Alice -- minimum abbreviation
-
- PREREQUISITES
-
- Nnodes
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-019 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- OUTFILE
-
- SYNTAX
-
- OUtfile [=] (file_name | TERMINAL)
-
- DESCRIPTION
-
- Outfile is used to open or close a file for output reports. Reports
- will automatically be displayed at the terminal if the current echo
- setting is yes (the default). If the terminal keyword is specified
- and an outfile file is already open, the file will be closed. This
- command will destroy the contents of a file that exists previously.
-
- EXAMPLES
-
- outfile = my_file -- opens the file, my_reports
- outfile terminal -- closes the file
- ou reports -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-020 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- PAGING
-
- SYNTAX
-
- PAging (Yes | No)
-
- DESCRIPTION
-
- Paging allows the terminal to be treated like a line printer by
- putting in new page marks. This is useful if run in batch, since
- the session input and output reports may be going to the same file.
- Reports generally start with a new page mark. 'Paging No' is the
- default.
-
- EXAMPLES
-
- paging yes
-
- pa n -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-021 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- PBRANCH
-
- SYNTAX
-
- PBranch [=] {probability_value}
-
- DESCRIPTION
-
- Pbranch is used to specify the branching probabilities between nodes
- in a network layout. Each node has associated with it a set of
- probabilities that a job terminating at the node will continue on
- to another node. There are exactly as many probabilities associated
- with each node as there are nodes. The probabilites for each node
- must sum to 1.
-
- EXAMPLES
-
- node = a node b
- pbranch = .3, .2 0.5 pb .6 .3 .1 -- minimum abbrev
-
- PREREQUISITES
-
- Nnodes, order, njobs, node.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-022 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- PROMPT
-
- SYNTAX
-
- PRompt
-
- DESCRIPTION
-
- Prompt is used to enter a mode where specific questions are asked
- by the program in order to build a complete model. There is no
- way to exit this mode short of entering a complete model definition.
- This command will cause a model reset upon user confirmation. The
- prompt sign is 'P>' as opposed to 'E>' in the normal editing mode.
-
- EXAMPLES
-
- E> prompt
- Enter the number of nodes.
- P> 3
- etc.
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-023 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- QUIT
-
- SYNTAX
-
- Quit
-
- DESCRIPTION
-
- Quit is used to end a session. The user must verify this action as
- a precaution against not entering the save command.
-
- EXAMPLE
-
- E> quit -- could use 'q' as well.
-
- This will cause the current model definition to be lost.
- Continue (Yes, No)?
- P> no -- could use 'n' as well
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-024 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- RATES
-
- SYNTAX
-
- RAtes [=] {real_value}
-
- DESCRIPTION
-
- Rates is used to specify the rate parameter for an exponential or
- erlang distribution and the rate parameter(s) for a coxian
- distribution.
-
- EXAMPLES
-
- node = a_register node b_register
- dist = coxian dist exponential
- nstages = 3 ra 5.4 -- minimum abbreviation
- rates = 1.5, 2.5 3.5
-
- PREREQUISITES
-
- Nnodes, order, njobs, node, distribution. If coxian, nstages.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-025 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- REPORT
-
- SYNTAX
-
- REPort (Arrival_Frequencies | SERV_Times | REsponse_times |
- Qlength_distributions | Throughput) node_list
-
- REPort (ROuting | SERV_Requirements) [FROM] node_list
- [TO] node_list [BY] node_list
-
- REPort (Model | Normalizations | Pbranch)
-
- node_list ::= (node_name | ({node_name}) | ALL
-
- DESCRIPTION
-
- Report generates output reports based on the latest model definition
- that was run. A summary of the reports follows.
-
- Arrival_frequencies:
- For each node in node_list, reports number of servers, discipline,
- and relative arrival frequency.
- Model:
- Reports the model definition consisting of edit commands.
- Normalizations:
- Reports normalizations constants for all values from 0 to the
- number of jobs in the model.
- Pbranch:
- Reports branching probabilities from all nodes to all nodes.
- Qlength_distributions:
- For each node in node_list and each njobs value from 0 to
- the number of jobs, reports the probability that the node has
- that many jobs.
- Response_times:
- Reports the response time mean, response time variance, and
- coefficient of variation for each node in node_list.
- Routing:
- Reports service tour mean and variance for each combination of
- nodes, taking one from each of the 'from', 'to' and 'by' node
- lists.
- Serv_requirements:
- Reports the service requirement mean, service requirement
- variance, and the residence time mean for each combination of
- nodes, taking one from each of the 'from', 'to' and 'by' node
- lists.
- Serv_times:
- For each node in node_list, reports distribution, mean service
- time, service time variance, coefficient of variation, service
- rate(s), number of stages, cbranch probabilities.
- Throughput:
- For each node in node_list, reports mean queue length, queue
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-026 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- length variance, coefficient of variation, throughput, and
- utilization.
-
- EXAMPLES
-
- report arrival_frequencies (a b c)
- rep t all -- abbreviations
- report routing from (a b) to all by c
- rep serv_r a (a b) all -- abbreviations
- report model
- rep p -- abbreviations
-
- PREREQUISITES
-
- Run
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-027 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- RESET
-
- SYNTAX
-
- REset
-
- DESCRIPTION
-
- Reset has the effect of undoing all model definition commands (those
- except for echo, outfile, paging). The user must verify this action
- as a precaution against not entering the save command.
-
- EXAMPLES
-
- E> reset -- could use 're' as well
-
- This will cause the current model definition to be lost.
- Continue (Yes, No)?
- P> yes -- could use 'y' as well
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-028 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- RUN
-
- SYNTAX
-
- RUn
-
- DESCRIPTION
-
- Run is used to execute the closed model calculations on the current
- model definition. All input parameters are checked before the
- calculations are performed.
-
- EXAMPLE
-
- run
-
- PREREQUISITES
-
- The model must be complete.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-029 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- SAVE
-
- SYNTAX
-
- SAve [=] file_name
-
- DESCRIPTION
-
- Save is used to save the current model definition onto a file,
- whether it be complete or not. This command will destroy the
- contents of file_name if it already exists. The commands that are
- saved include title, nnodes, order, njobs, node, pbranch,
- discipline, distribution, rates, nstages, cbranch, and endnode.
-
- EXAMPLES
-
- save my_file
-
- sa your_file -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-030 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- SHOW
-
- SYNTAX
-
- SHow (Title | NNodes | Order | NJobs | NOde node_name | Model)
-
- DESCRIPTION
-
- This command is used to display the current model definition
- parameters. 'Show model' will display all of them.
-
- EXAMPLES
-
- E> show title
- title = Assembly Line
-
- E> sh m -- minimum abbreviation, will show entire model
- nnodes = 3
- .
- .
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-031 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- TITLE
-
- SYNTAX
-
- Title [=] title_specification
-
- DESCRIPTION
-
- Title is used to a specify a title for the model that will appear
- on all output reports. The title specification must appear on the
- same line as the command and cannot exceed 50 characters in
- length.
-
- EXAMPLES
-
- title My Main Model
-
- t A vERy OdD tITle -- minimum abbreviation
-
- PREREQUISITES
-
- None
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 8-032 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 9. Appendix B - Closed Network Model Class
-
-
-
- Symbol Definition
-
-
- $M$ number of jobs in the network
-
- $N$ number of nodes in the network
-
- $P$ matrix of connectivity, irreducible stochastic matrix
- of order $N$
-
- $p probability of a job branching from node $i$ to node
- $j$, for $ 1 <= i,j <= N$
-
- $bold number of jobs at node $i$ in steady-state
-
- $s number of servers at node i
-
- $ service or processing time at node $i$
-
- $b $E [ ( bold S sub i ) sup r ] r = 1, 2, ... $
-
- $bold response time at node $i$ in steady-state
-
- $mu service rate at node i, $mu sub i = b sub i1 sup -1$
-
-
- 9.1 Service Distributions
-
- The following service distributions are available for
- selection for each node.
-
-
- 9.1.1 Exponential Distribution
-
- where $mu$ is the specified service rate parameter.
-
-
- 9.1.2 R-Stage Erlang Distribution
-
- where $mu$ is the specified service rate per stage and $R$ is the
- specified number of stages.
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 9-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 9.1.3 R-Stage Coxian Distribution
-
- The $r sup th$ service moment is given by the first element
- of the vector
-
-
-
- where $R$ is the specified number of stages, $I$ is the identity
- matrix of order $R$, $P sup *$ and $D sup (r)$ are square
- matrices of order $R$, and $b sup (r)$ and $delta sup (r)$ is a
- column vector of dimension $R$. The following are expressions
- for $P sup *$, $delta sup (r)$, and $D sup (n)$:
-
-
-
- where $p sub i$ is the specified probability of a job continueing
- to the $(i+1) sup th$ stage.
-
-
-
- where $mu sub i$ is the service rate at stage $i$, $p sub i + q
- sub i = 1$, and $q sub R = 1$,
-
-
-
- and $b sup (0)$ is a column vector of dimension $R$,
-
-
-
-
- 9.2 Closed Model Equations
-
- $alpha vec = ( bold alpha sub 1 , bold alpha sub 2 , ...,
- bold alpha sub N )$ is defined as the steady state flow per unit
- time through a node and is any positive row vector which
- satisfies the matrix equation
-
-
-
-
-
-
-
-
-
- $bold c vec = ( c sub 1 , c sub 2 , ..., c sub N )$ is defined
- as a state of the network system where $c sub i$ is the number of
- jobs at node $i$, and
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 9-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- $bold psi ( bold c vec )$ is defined as the equilibrium or
- steady-state probability of the network being in state $bold c
- vec$. Now
-
-
-
- where
-
-
-
- and
-
-
-
- with
-
-
-
- $bold phi sub i ( m, x )$ is defined as the marginal probability
- of $m$ jobs at node $i$.
-
-
-
- where
-
-
-
- $bold Q sub i ( M )$ is the number of jobs at node $i$ in
- steady-state.
-
-
-
-
-
- $bold U sub i ( M )$ is the steady-state probability that station
- $i$ is active, or more commonly called the utilization, and is
- given by
-
-
-
- If node $i$ has the PS or PR-LCFS discipline the first moment of
- the response time $bold T sub i$ is
-
-
-
- where $tau sub i ( M )$ is the throughput at node $i$, and is
- given by
-
-
-
- If node $i$ has the NQ discipline then
-
-
- Item 0037 Ford Aerospace &
- June 1985 9-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
-
- If node $i$ has the FCFS queue discipline the first and second
- moments of the response time $bold T sub i$ are given by the next
- two equations, where $mu sub bold i$ is the service rate at node
- $i$.
-
-
-
-
-
- For any queueing discipline at node $i$
-
-
- 9.2.1 Equations for Routing Behavior and Return Times
-
-
- $bold P under ( k )$ is formed from $bold P$ by replacing
- the $k sup th$ column with $0$'s.
-
-
-
- $m sub ik (m sub i,k sup * )$ is the mean (variance) of the
- number of service tours required to return to node $k$ given that
- the job starts from node $i$.
-
-
-
-
-
- $M sub ij sup k ( V sub ij sup k )$ is the mean (variance) of
- the number of visits by a job to node $j$, before the job returns
- to node $k$, given that the job starts from node $i$.
-
-
-
-
-
- $S sub ij sup k ( W sub ij sup k )$ is the mean (variance) of
- the total service requirement of a job at node $j$ before the job
- returns to node $k$, given that the job starts from node $i$. $R
- sub ij sup k$ is the mean residence time of a job at node $j$
- before the job returns to node $k$, given that the job starts
- from node $i$.
-
-
-
-
-
-
-
- $R sub ik sup *$ is the mean time required, and $S sub ik sup *$
-
-
- Item 0037 Ford Aerospace &
- June 1985 9-004 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- is the mean service requirement, for a job to return to node $k$
- given that the job starts from node $i$.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 9-005 Communications Corporation
-
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 10. Appendix C - Test Cases
-
-
-
- TEST CASE 1
-
-
-
-
-
- <<< Test Case 1 >>> 85/7/2
- Model Definition 12:14
-
- Title = <<< Test Case 1 >>>
- Nnodes = 3
- Order = NODE1 NODE2 NODE3
- Njobs = 5
- Node = NODE1
- Pbranch = 0.4 0.3 0.3
- Discipline = FCFS
- Nservers = 1
- Distribution = EXPONENTIAL
- Rate = 3.0
- Endnode
- Node = NODE2
- Pbranch = 0.2 0.6 0.2
- Discipline = FCFS
- Nservers = 2
- Distribution = EXPONENTIAL
- Rate = 2.0
- Endnode
- Node = NODE3
- Pbranch = 0.2 0.2 0.6
- Discipline = FCFS
- Nservers = 3
- Distribution = EXPONENTIAL
- Rate = 1.0
- Endnode
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-001 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Branch Probabilities 12:14
-
- -------------------------------------------------------------------------
-
-
- LEGEND
-
- 1=NODE1 2=NODE2 3=NODE3
- -------------------------------------------------------------------------
-
-
- NODE1
- 1) .4 2) .3 3) .3
- NODE2
- 1) .2 2) .6 3) .2
- NODE3
- 1) .2 2) .2 3) .6
-
-
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Relative Arrival Frequencies 12:14
-
- Node Number of Queue Relative Arrival
- Name Servers Discipline Frequency
- _____________________________________________________________________
-
- NODE1 1 FCFS 0.25
- NODE2 2 FCFS 0.375
- NODE3 3 FCFS 0.375
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-002 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Service Time 12:14
-
- Node Service Mean Service Service Time Coefficient of
- Name Distribution Time Variance Variation
- _______________________________________________________________________________
-
- NODE1 EXPONENTIAL 0.3333 0.1111 1.0
-
- Service Rate: 3.0
-
-
- NODE2 EXPONENTIAL 0.5 0.25 1.0
-
- Service Rate: 2.0
-
-
- NODE3 EXPONENTIAL 1.0 1.0 1.0
-
- Service Rate: 1.0
-
-
-
-
-
-
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Queue Lengths 12:14
-
- Node Mean Queue Queue Length Coefficient of Throughput Utilization
- Name Length Variance Variation
- _______________________________________________________________________________
-
- NODE1 0.8617 1.091 1.212 1.545 0.5151
- NODE2 1.416 1.417 0.8407 2.318 0.7507
- NODE3 2.722 1.769 0.4886 2.318 0.9562
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-003 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Response Times 12:15
-
- Node Mean Response Response Time Coefficient of
- Name Time Variance Variation
- _________________________________________________________________
-
- NODE1 0.5576 0.2704 0.9326
- NODE2 0.611 0.3106 0.9122
- NODE3 1.174 1.114 0.8986
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Queue Length Distributions 12:15
-
- m P {Q = m} Node: NODE1
- ________________________
-
- 0 0.4849
- 1 0.2801
- 2 0.1483
- 3 0.06471
- 4 0.01917
- 5 2.84e-3
-
-
- m P {Q = m} Node: NODE2
- ________________________
-
- 0 0.2493
- 1 0.3425
- 2 0.2215
- 3 0.1264
- 4 0.05005
- 5 0.01024
-
-
- m P {Q = m} Node: NODE3
- ________________________
-
- 0 0.04384
- 1 0.1512
- 2 0.2482
- 3 0.2494
- 4 0.2103
- 5 0.09706
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-004 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Routing Behavior 12:15
-
- Start At Node: NODE1 Return To Node: NODE1
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
- NODE2 1.5 6.25
- NODE3 1.5 6.25
-
- Total 4.0 34.0
-
- Start At Node: NODE1 Return To Node: NODE2
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
- NODE1 2.222 2.716
- NODE3 1.667 6.667
-
- Total 3.889 28.21
-
- Start At Node: NODE1 Return To Node: NODE3
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
- NODE1 2.222 2.716
- NODE2 1.667 6.667
-
- Total 3.889 28.21
-
- Start At Node: NODE2 Return To Node: NODE1
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
- NODE2 3.333 7.778
- NODE3 1.667 6.667
-
- Total 5.0 45.0
-
- Start At Node: NODE2 Return To Node: NODE2
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
- NODE1 0.6667 1.852
- NODE3 1.0 4.667
-
- Total 2.667 16.74
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-005 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Routing Behavior 12:15
-
- Start At Node: NODE2 Return To Node: NODE3
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE1 1.111 2.593
- NODE2 3.333 7.778
-
- Total 4.444 33.83
-
- Start At Node: NODE3 Return To Node: NODE1
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE2 1.667 6.667
- NODE3 3.333 7.778
-
- Total 5.0 45.0
-
- Start At Node: NODE3 Return To Node: NODE2
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE1 1.111 2.593
- NODE3 3.333 7.778
-
- Total 4.444 33.83
-
- Start At Node: NODE3 Return To Node: NODE3
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE1 0.6667 1.852
- NODE2 1.0 4.667
-
- Total 2.667 16.74
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-006 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Service Requirements 12:15
-
- Start At Node: NODE1 Return To Node: NODE1
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
- NODE2 0.75 1.937 0.9164
- NODE3 1.5 7.75 1.762
-
- Total 2.583 3.236
-
- Start At Node: NODE1 Return To Node: NODE2
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
- NODE1 0.7407 0.5487 1.239
- NODE3 1.667 8.333 1.957
-
- Total 2.407 3.197
-
- Start At Node: NODE1 Return To Node: NODE3
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
- NODE1 0.7407 0.5487 1.239
- NODE2 0.8333 2.083 1.018
-
- Total 1.574 2.257
-
- Start At Node: NODE2 Return To Node: NODE1
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
- NODE2 1.667 2.778 2.037
- NODE3 1.667 8.333 1.957
-
- Total 3.333 3.994
-
- Start At Node: NODE2 Return To Node: NODE2
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
- NODE1 0.2222 0.2798 0.3718
- NODE3 1.0 5.667 1.174
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-007 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Total 1.722 2.157
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-008 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 1 >>> 85/7/2
- Service Requirements 12:16
-
- Start At Node: NODE2 Return To Node: NODE3
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE1 0.3704 0.4115 0.6196
- NODE2 1.667 2.778 2.037
-
- Total 2.037 2.656
-
- Start At Node: NODE3 Return To Node: NODE1
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE2 0.8333 2.083 1.018
- NODE3 3.333 11.11 3.915
-
- Total 4.167 4.933
-
- Start At Node: NODE3 Return To Node: NODE2
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE1 0.3704 0.4115 0.6196
- NODE3 3.333 11.11 3.915
-
- Total 3.704 4.535
-
- Start At Node: NODE3 Return To Node: NODE3
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE1 0.2222 0.2798 0.3718
- NODE2 0.5 1.417 0.611
-
- Total 1.722 2.157
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-009 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
-
-
-
- TEST CASE 2
-
-
-
-
-
- <<< Test Case 2 >>> 85/7/2
- Model Definition 12:17
-
- Title = <<< Test Case 2 >>>
- Nnodes = 4
- Order = NODE_A NODE_B NODE_C NODE_D
- Njobs = 25
- Node = NODE_A
- Pbranch = 0.3 0.3 0.3 0.1
- Discipline = FCFS
- Nservers = 1
- Distribution = EXPONENTIAL
- Rate = 4.0
- Endnode
- Node = NODE_B
- Pbranch = 0.2 0.3 0.3 0.2
- Discipline = P_SHARE
- Nservers = 1
- Distribution = EXPONENTIAL
- Rate = 3.0
- Endnode
- Node = NODE_C
- Pbranch = 0.4 0.0 0.4 0.2
- Discipline = PR_LCFS
- Nservers = 1
- Distribution = ERLANG
- Nstages = 2
- Rate = 2.0
- Endnode
- Node = NODE_D
- Pbranch = 0.5 0.0 0.5 0.0
- Discipline = NQ
- Nservers = 25
- Distribution = COXIAN
- Nstages = 3
- Cbranch = 0.9 0.75
- Rates = 1.0 1.0 1.0
- Endnode
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-010 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Branch Probabilities 12:17
-
- -------------------------------------------------------------------------
-
-
- LEGEND
-
- 1=NODE_A 2=NODE_B 3=NODE_C 4=NODE_D
-
- -------------------------------------------------------------------------
-
-
- NODE_A
- 1) .3 2) .3 3) .3 4) .1
- NODE_B
- 1) .2 2) .3 3) .3 4) .2
- NODE_C
- 1) .4 2) 0.0 3) .4 4) .2
- NODE_D
- 1) .5 2) 0.0 3) .5 4) 0.0
-
-
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Relative Arrival Frequencies 12:17
-
- Node Number of Queue Relative Arrival
- Name Servers Discipline Frequency
- _____________________________________________________________________
-
- NODE_A 1 FCFS 0.349
- NODE_B 1 P_SHARE 0.1496
- NODE_C 1 PR_LCFS 0.3639
- NODE_D 25 NQ 0.1376
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-011 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Service Time 12:17
-
- Node Service Mean Service Service Time Coefficient of
- Name Distribution Time Variance Variation
- _______________________________________________________________________________
-
- NODE_A EXPONENTIAL 0.25 0.0625 1.0
-
- Service Rate: 4.0
-
-
- NODE_B EXPONENTIAL 0.3333 0.1111 1.0
-
- Service Rate: 3.0
-
-
- NODE_C ERLANG 1.0 0.5 0.7071
-
- Number of Stages: 2
- Service Rate : 2.0
-
-
- NODE_D COXIAN 2.575 3.019 0.6748
-
- Number of Stages: 3
- Service Rates : 1.0 1.0 1.0
- Cbranch Probs : .9 .75
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-012 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Queue Lengths 12:17
-
- Node Mean Queue Queue Length Coefficient of Throughput Utilization
- Name Length Variance Variation
- _______________________________________________________________________________
-
- NODE_A 0.3153 0.4147 2.042 0.9589 0.2397
- NODE_B 0.1587 0.1839 2.702 0.411 0.137
- NODE_C 23.55 1.574 0.05326 1.0 1.0
- NODE_D 0.9736 0.9736 1.013 0.3781 0.6223
-
-
-
-
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Response Times 12:17
-
- Node Mean Response Response Time Coefficient of
- Name Time Variance Variation
- _________________________________________________________________
-
- NODE_A 0.3288 0.1081 1.0
- NODE_B 0.3862
- NODE_C 23.55
- NODE_D 2.575 3.019 0.6748
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-013 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Queue Length Distributions 12:18
-
- m P {Q = m} Node: NODE_A
- ________________________
-
- 0 0.7603
- 1 0.1823
- 2 0.04369
- 3 0.01047
- 4 2.511e-3
- 5 6.019e-4
- 6 1.443e-4
- 7 3.459e-5
- 8 8.293e-6
- 9 1.988e-6
- 10 4.766e-7
- 11 1.142e-7
- 12 2.739e-8
- 13 6.566e-9
- 14 1.574e-9
- 15 3.773e-10
- 16 0.0
- 17 0.0
- 18 0.0
- 19 0.0
- 20 0.0
- 21 0.0
- 22 0.0
- 23 0.0
- 24 0.0
- 25 0.0
-
-
- m P {Q = m} Node: NODE_B
- ________________________
-
- 0 0.863
- 1 0.1182
- 2 0.01619
- 3 2.218e-3
- 4 3.039e-4
- 5 4.163e-5
- 6 5.703e-6
- 7 7.812e-7
- 8 1.07e-7
- 9 1.466e-8
- 10 2.008e-9
- 11 2.751e-10
- 12 0.0
- 13 0.0
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-014 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 14 0.0
- 15 0.0
- 16 0.0
- 17 0.0
- 18 0.0
- 19 0.0
- 20 0.0
- 21 0.0
- 22 0.0
-
-
-
-
-
- Page 2 <<< Test Case 2 >>> 85/7/2
- Queue Length Distributions 12:18
-
- m P {Q = m} Node: NODE_B
- ________________________
-
- 23 0.0
- 24 0.0
- 25 0.0
-
-
- m P {Q = m} Node: NODE_C
- ________________________
-
- 0 9.022e-8
- 1 1.048e-7
- 2 1.216e-7
- 3 0.0
- 4 0.0
- 5 0.0
- 6 0.0
- 7 0.0
- 8 0.0
- 9 3.462e-7
- 10 4.02e-7
- 11 9.337e-7
- 12 1.084e-6
- 13 1.259e-6
- 14 5.116e-6
- 15 1.697e-5
- 16 8.13e-5
- 17 3.218e-4
- 18 1.231e-3
- 19 4.468e-3
- 20 0.01502
- 21 0.04517
- 22 0.1158
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-015 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- 23 0.2354
- 24 0.3347
- 25 0.2478
-
-
- m P {Q = m} Node: NODE_D
- ________________________
-
- 0 0.3777
- 1 0.3677
- 2 0.179
- 3 0.05809
- 4 0.01414
- 5 2.753e-3
- 6 4.467e-4
- 7 6.213e-5
- 8 7.561e-6
- 9 8.179e-7
- 10 7.963e-8
- 11 7.048e-9
- 12 5.718e-10
- 13 0.0
- 14 0.0
-
-
-
-
-
- Page 3 <<< Test Case 2 >>> 85/7/2
- Queue Length Distributions 12:18
-
- m P {Q = m} Node: NODE_D
- ________________________
-
- 15 0.0
- 16 0.0
- 17 0.0
- 18 0.0
- 19 0.0
- 20 0.0
- 21 0.0
- 22 0.0
- 23 0.0
- 24 0.0
- 25 0.0
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-016 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Routing Behavior 12:18
-
- Start At Node: NODE_A Return To Node: NODE_A
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE_B 0.4286 0.6122
- NODE_C 1.043 2.041
- NODE_D 0.3943 0.3965
-
- Total 2.866 12.25
-
- Start At Node: NODE_A Return To Node: NODE_B
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE_A 3.333 7.778
- NODE_C 2.333 12.44
- NODE_D 0.8 1.76
-
- Total 6.467 91.89
-
- Start At Node: NODE_A Return To Node: NODE_C
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE_A 1.918 1.76
- NODE_B 0.8219 1.43
- NODE_D 0.3562 0.3562
-
- Total 3.096 15.59
-
- Start At Node: NODE_A Return To Node: NODE_D
-
- Node Service Tour Mean Service Tour Variance
- _____________________________________________________________
-
- NODE_A 3.043 6.219
- NODE_B 1.304 2.665
- NODE_C 2.174 6.648
-
- Total 6.522 74.64
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-017 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
- Page 1 <<< Test Case 2 >>> 85/7/2
- Service Requirements 12:19
-
- Start At Node: NODE_A Return To Node: NODE_A
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE_B 0.1429 0.1156 0.1655
- NODE_C 1.043 2.562 24.56
- NODE_D 1.015 3.82 1.015
-
- Total 2.451 26.07
-
- Start At Node: NODE_A Return To Node: NODE_B
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE_A 0.8333 0.6944 1.096
- NODE_C 2.333 13.61 54.96
- NODE_D 2.06 14.09 2.06
-
- Total 5.227 58.11
-
- Start At Node: NODE_A Return To Node: NODE_C
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE_A 0.4795 0.2299 0.6306
- NODE_B 0.274 0.2502 0.3175
- NODE_D 0.9171 3.437 0.9171
-
- Total 1.671 1.865
-
- Start At Node: NODE_A Return To Node: NODE_D
-
- Node Service Requirement Service Requirement Residence Time
- Name Mean Variance Mean
- ___________________________________________________________________________
-
- NODE_A 0.7609 0.5789 1.001
- NODE_B 0.4348 0.4411 0.5038
- NODE_C 2.174 7.735 51.2
-
- Total 3.37 52.71
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-018 Communications Corporation
-
-
-
-
-
-
-
- Queue Modeling Tool
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- Item 0037 Ford Aerospace &
- June 1985 10-019 Communications Corporation
-
-
-
-
-
-
-
- CONTENTS
-
-
- 1. Introduction 1-001
-
- 2. Computer Program System Capabilities 2-001
- 2.1 Purpose 2-001
- 2.2 General Description 2-001
- 2.3 Functions Performed 2-002
-
- 3. Notification of Future Enhancements 3-001
-
- 4. Installing Queue Modeling Tool 4-001
- 4.1 Deliverable Items 4-001
- 4.2 Compiling, Linking, and Executing the Tool 4-001
- 4.3 Math library installation 4-002
- 4.4 Linking the Tool 4-003
- 4.5 Installing The Help Facility 4-004
- 4.6 Exceptions 4-005
- 4.7 Compilation Units 4-006
- 4.8 Compilation Order 4-007
-
- 5. State of Art Uses of Ada 5-001
- 5.1 Ada Features 5-001
- 5.2 Naming Conventions 5-001
- 5.3 Pretty_Printing 5-002
-
- 6. Description of the Closed Queueing Model 6-001
-
- 7. Output Reports Produced 7-001
- 7.1 Model Inputted by the User 7-001
- 7.2 Queue Lengths 7-002
- 7.3 Response Times 7-002
- 7.4 Normalization Constants 7-003
- 7.5 Queue Length Distributions 7-003
- 7.6 Routing Behavior 7-003
- 7.7 Node Service Requirements and Return Times 7-003
-
- 8. Appendix A - Description of Commands 8-001
-
- 9. Appendix B - Closed Network Model Class 9-001
- 9.1 Service Distributions 9-001
- 9.2 Closed Model Equations 9-002
-
- 10. Appendix C - Test Cases 10-001
-
- - i -
-