HPAVC

home *** CD-ROM | disk | FTP | other *** search

/ HPAVC / HPAVC CD-ROM.iso / pc / GAF200B.ZIP / DOC / FCL.DOC < prev next >

Wrap

Text File | 1994-10-29 | 28.7 KB | 737 lines

Introduction This document explains how to use FCL and build a control simulation under GAF. Following topics are covered in this document: - brief description of basic concept - simple steps to build a simulation - FCL reference manual A segment, the basic module in GAF, is a text file contains these information for (or related to) a segment. - configuration information - variable declarations (IN, OUT, INOUT, and LOCAL) - symbol (fuzzy membership set) declarations - enable condition and cycle time of the segment - initialization - reset - pre-processing - post-processing - fuzzy rules - data table(s) A typical FCL segment layout will look like this. FCL construct Refer to Section ------------------------------- ------------------ configuration_declaration Configuration segment_declaration Declare segment cycle_declaration Cycle Time variable_declaration(s) Declare variable enable_declaration Enable condition symbol_declaration(s) Declare fuzzy symbol preset_declaration(s) Preset init_declaration Initialization reset_declaration Reset pre_processing Pre-processing post_processing Post-processing rule_declaration(s) Define rule data_declaration(s) Data Table end_declaration Declare segment Just like dividing a control system into several manageable portions, GAF allows a control system to be divided into segments. For example, a simulated control environment under GAF may have these different segments: - control segment(s) - reference calculation and supporting segment(s) - simulation (or feedback) segment(s) - evaluation segment The feedback and evaluation segments are used for simulation and adaptation. Segment Construction To construct a segment is quite similar to construct a simple program. The followings are simple steps to construct a GAF segment. 1. Define all variables for this segment: All variables must be declared before they can be used either by fuzzy rules or any statements in this segment. There are four types of variables: IN, OUT, INOUT, and LOCAL. IN variables IN variables are used as "references" in this segment. IN variables must be in one of the categories: - calculated by other segment (as OUT) - an input from real world, i.e. a sensor, this input may be an OUT from a feedback segment in a simulation. - application constants, can be modified through interactive interface during simulation OUT variables OUT variables are calculated outputs by this segment. OUT variables are either: - an output signal to a real world device - be used by other segment as IN variables - for diagnostic purpose - evaluation result INOUT variables INOUT variables are both referenced and modified by this segment, i.e., it is the combination of both IN and OUT variables. Normally INOUT variables are referenced by this segment and other segments. All of the IN, OUT, and INOUT variables are visible to other segments, i.e. other segments can use this segment's OUT as its IN. LOCAL variables LOCAL variables are local to this segment. Normally, local variables are used to store temporary values so it can be used by fuzzy rules or other statements within the segment. 2. Fuzzy membership set declaration. The second step would be define the fuzzy membership set. GAF uses four points for each set. In general, fuzzy sets should be defined for each variable that is used in the fuzzy rules. For example, as described in the following figure. A simple way to define membership sets for a variable is to equaly divide the range of the variable into six to obtain five sets - very low (VL), low (LO), medium (M), high (HI), and very high (VH). ----- - - - ----- \ / \ / \ / \ / \ / \ / \ / \ / X X X X / \ / \ / \ / \ / \ / \ / \ / \ +-----+-----+-----+-----+-----+-----+ -1 -0.67 -0.33 0 0.33 0.67 1 | very low | medium | very high | | low | high | As the above diagram shows these five membership sets are symmatrically divided. After determine each membership set, then, use these declarations for the starting point. SYMBOL very_low (-1.00, -1.00, -0.67, -0.33, 1.0) SYMBOL low (-0.67, -0.33, -0.33 0, 1.0) SYMBOL medium (-0.33, 0, 0, 0.33, 1.0) SYMBOL high ( 0, 0.33, 0.33, 0.67, 1.0) SYMBOL very_high ( 0.33, 0.67, 1, 1, 1.0) 3. Define fuzzy rules. Defining fuzzy rules is straight forward. For example, based on common sense to control a water heater, a simple rule may be like: "if water usage is heavy and water is cold, then the gas fuel valve to heat the tank must be full open" This idea can turn into a rule like: IF water_usage IS heavy AND water IS cold THEN fuel_valve IS full_open Where water_usage and water are IN variables, fuel_valve is an OUT variable; heavy, cold, and full_open are membership sets defined for corresponding variables. Note that all variables that are part of "THEN" statement must be declared as OUT or INOUT. Instead of defining fuzzy rule, data set can be used to replace fuzzy rules. Please refer to step 8 for more details. 4. Initializes the segment. The initialization of a segment is to initialize all outputs (OUT and INOUT variables) and its local variables. The initialization will be executed when the system "comes up" (i.e., the first time it runs the segment) or by the request from interactive interface. All initialization formulas must be stated like this: what's in the text remarks ------------------- ----------------------------- INITIALIZATION must have this key word ... ; statements for initialization each statement must end with ';' RESET optional key word for ... ; reset processing statements PRE_PROCESSING optional key word for ... ; pre-processing statements POST_PROCESSING optional key word for ... ; post-processing statements END; must end this key word (with ';') For example, for the above water heat control, the segment should initialize the fuel_valve to close. INITIALIZATION fuel_valve = 0; PRE_PROCESSING ... END; 5. Reset processing Similar to initialization, the reset processing is to reset the segment when the segment is disabled. The reset processing is only necessary when it is different from the initialization. 6. Pre-processing and post-processing The pre-processing is a set of statements to be calculated before the fuzzy rules are inferenced. The post-processing is a set of statements to be calculated after the fuzzy rules are inferenced. The formulas can be used for calculating references or other simple mathematical models. Pre-processing is calculated before any fuzzy rule been evaluated, so the results of these formulas may be used by fuzyy rules to inference the result. Post-processing is calculated after the inference of fuzzy rules. 7. Segment enable condition and its cycle time The enable condition determines whether the segment shoud run or not. If enable condition is not specified GAF treats the segment as enabled all the time. The cycle time determines how frequently the segment will run, cycle time can be adjusted during simulation. 8. Build data table An alternative of defining fuzzy rules is to use data set. Whether it is measured data from the field or data generated from a modeling program, the data set can be plugged into GAF. Data set must be arranged into spread sheet like table, and using the key DATA_TABLE to indicate the contents and the order of data in the table. Note that the data set must be a complete data set, i.e., it must cover the entire range of each variable. Format Covention Conventions used for describing FCL syntax (format): - all key words are in upper case - key words are not case sensitive, but must be the same - options are enclosed within breckets [] - < | | > indicates selection of one out of two or more items Comment A comment can be anywhere in the segment, it starts with '!' and lasts to the end of the line. Declare Segment: segment_declaration, end_declaration There are three types of segments: normal control segment, feedback segment, and evaluation segment. Format for segment_declaration: FORMAT < SEGMENT | FEEDBACK | EMUL_SEGMENT | EVAL_SEGMENT > seg_name . . . seg_contents . . . END seg_name ; Where seg_name : the name of the segment seg_contents are statements for this segment Example SEGMENT main_control CYCLE_TIME 0.1 . . . END main_control; As shown in the FCL construct, a segment is enclosed within two keys. The first key can be one of the followings, followed by the name of the segment. SEGMENT for normal (control) segment EMUL_SEGMENT & FEEDBACK for emulation/feedback segment EVAL_SEGMENT for adaptive evaluation segment A segment must always end with the END keyword followed by the name of the segment. Note that in simulation and adapt modes the feedback segment's output value will be feed back to the control segment(s) based on variable name(s). Cycle Time: cycle_declaration FORMAT CYCLE_TIME value Where value : the cycle time in seconds Example SEGMENT main_control CYCLE_TIME 0.1 . . . END main_control; Cycle_time defines how frequently the segment runs. For example a value of 0.1 defines the segment runs every 100 milliseconds. Note that actual delay time between two consecutive runs may be longer than 100 milliseconds if GAF is heavily loaded with segments. Declare Variable: variable_declaration There are four different kinds of variables: IN, OUT, INOUT, and LOCAL. The segment will only read (reference) from those IN variables, and write (generate new value) to the OUT variables. The segment can read and write to both INOUT and LOCAL variables, however, local variables are not visible from outside the segment. FORMAT var_type var_name (min_value, max_value) Where var_type : one of the 4 types IN, OUT, INOUT, LOCAL var_name : the variable name min_value : the minimum value of the variable max_value : the maximum value of the variable Example SEGMENT main_control CYCLE_TIME 0.1 IN Reference (0, 100) OUT Result (0, 10) LOCAL Temp (-1, 1) . . . END main_control; Variable name must be unique in the same segment. A variable should not be declared as an output variable (OUT or INOUT) in more than one segments. GAF Global Variable DELTA_TIME: The delta_time variable is the time gap between the current segment run time and the last time the segment ran. Delta time is in seconds (floating point), it can be used only in math processing. Example: Speed = ( Position - Last_position ) / Delta_time ; Declare Fuzzy Symbol: symbol_declaration The basic fuzzy membership set definition is a trapezoid with 4 points below, low, high, above, and a max_truth value. _____ <----- max_truth: maximum truth value / \ / \ / \ -------+--+-----+--+------- <----- 0: minimum truth value | | | | | | | above: false above | | high: high truth | low: low truth below: false below As depicted in the diagram above, the truth for different values are: if value <= below truth = 0 if value >= above truth = 0 if low <= value <= high truth = max_truth for other values use interpolation to derive truth The range of truth value is 0.0 to 1.0. FORMAT SYMBOL sym_name [OF var_name] fuzzy_membership_set Format for fuzzy_membership_set ( below, low, high, above [, truth [, center ] ] ) Where sym_name : the name of the symbol var_name : the name of the variable for this symbol fuzzy_membership_set : defines the four points of a membership set and truth : the max truth value center : the center value Example SEGMENT main_control CYCLE_TIME 0.1 IN Reference (0, 100) OUT Result (0, 10) LOCAL Temp (-1, 1) SYMBOL small OF Reference (0, 0, 10, 20, 1.0) SYMBOL strong (8, 9, 10, 10, 1, 10) . . . END main_control; The fuzzy symbol declaration is used to declare a named fuzzy membership set. This named symbol can then be used in the fuzzy rules. The symbol defines the shape of a trapezoid fuzzy membership set, as described above, and the optional truth value and the center value. If the center value is specified, GAF uses this value, otherwise GAF calculates the center of gravity as the center value. A symbol can be defined for all variables or defined for a specific variable with "OF var_name" option. If a symbol is defined for a specific variable, it can only be used for that variable. Symbol names are not required to be unique in a segment. For example, two symbols "VERY_SMALL" can be defined for two different variables "SPEED" and "POPULATION" with two different shapes as in the following example. IN Speed (-100, 100) IN Population (0, 1000) SYMBOL VerySmall OF Speed (-5, 0, 0, 5) SYMBOL VerySmall OF Population (0, 50, 50, 100) Math Processing Statement Besides fuzzy rule, GAF supports non-fuzzy math calculation: the math statements. The math statements are very similar to high level languages such as C++ and Ada. Math statements can be further divided into two categories: condition statement and assignment statement. Assignment statement is used to assign a variable with a value or the result of a math formula. The format for an assignment statement is: FORMAT var_name = [ unary_op ] math_val [ binary_op math_val ] ; Where var_name: the name of the target variable math_val: a constant or name of a variable unary_op: unary operator: - binary_op: binary operator: + - * / The condition statement is used to direct GAF to execute different assignments based on non-fuzzy boolean logic. The format for a condition statement is: FORMAT IF logic_expression THEN true_statements [ ELSE false_statements ] ENDIF; Where - logic expression is a math formula, if the result is TRUE then the true assignment statements will be executed. If the result is FALSE, then the false assignment statements will be executed. Available logical operators are: > greater than < less than >= greater or equal <= less or equal == equal <> not equal && and operator AND and operator || or operator OR or operator - true_statement and false_statement are assignment statements as described. Note that math statements are non-fuzzy statements (for current version). Since bivalent logic is simply an extreme of a fuzzy set, GAF does not support "boolean" variable. However, boolean logic can be emulated by fuzzy variable as in this example. LOCAL tmp_bool (0, 1) LOCAL tmp_val (0, 10) IF tmp_bool THEN ! use tmp_bool as a boolean tmp_val = 3; ENDIF; IF ( tmp_bool == 1 ) THEN ! compare tmp_bool with 1 tmp_val = 5; END_IF; IF ( tmp_val > 9 ) THEN tmp_val = 9; ENDIF; Initialization: init_declaration Reset: reset_declaration Pre-Processing: pre_processing Post-Processing: post_processing The initialization declaration defines how the segment will be initialized. The reset declaration defines how to reset a segment when the segment is disabled. Pre/Post processing declarations define the statements to be executed before/after fuzzy inference. FORMAT [ [ INITIALIZATION init_statement(s) ] [ RESET reset_statement(s) ] [ < PRE_PROCESSING | BEGIN > pre_statement(s) ] [ POST_PROCESSING post_statement(s) ] END; ] Where - init_statements are math statements for initializing the segment. - reset_statements are math statements to reset the segment. - pre_statements are math statements to be executed before the fuzzy rule inference for the segment. - post_statements are math statements to be executed after the fuzzy rule inference for the segment. Example INITIALIZATION speed = 0; PRE_PROCESSING IF position > 50 THEN speed = 3.5; ELSE speed = speed + 2; END IF; END; Preset: preset_declaration Just like the initialization and reset declarations, a preset declaration is a sequence of assignments to setup the segment to a known state. Normally this is done by assigning values (constants) to local and output variables. There is no limit of the number of presets in a segment. Presets are used for genetic adapt for testing new rules' result. They also can be requested by user during simulation and check modes. FORMAT: PRESET preset_statement(s) END; Example PRESET Reference = 100; Position = 0; END ; Define Rule: rule_declaration FCL uses IF THEN construct to define a fuzzy rule. The IF statement contains multiple fuzzy evaluations bound with AND or OR. The format of fuzzy evaluation is defined as: variable IS symbol which describes "What is the truth value of the current value of 'variable' in the fuzzy membership set 'symbol'?". The result of the evaluation will be the truth value between 0 and 1 as described in the symbol_declaration. The overall result of the IF statement will be the fuzzy AND/OR result (min/max of all truth values). The THEN statement contains only one fuzzy induction. The format of a fuzzy induction is one of these forms: variable IS symbol variable + symbol variable - symbol The variable must be an OUT, INOUT, or LOCAL variable. The first format simply defines the result of 'variable' is the 'symbol' fuzzy membership set with the max truth be the overall result from the IF statement. The second and third format is to increment/decrement the current 'variable' by the amount defined by 'symbol'. Example SEGMENT main_control CYCLE_TIME 0.1 IN Reference (0, 100) IN Speed (-1, 1) OUT Result (0, 10) LOCAL Temp (-1, 1) SYMBOL small OF Reference (0, 0, 10, 20, 1.0) SYMBOL slow OF Speed (0, 1, 2, 3, 1) SYMBOL strong (8, 9, 10, 10, 1, 10) IF Reference IS small AND Speed IS slow THEN Result IS strong . . . . END main_control; A rule can be repeated more than once in a segment to intensify the effect of the rule. Data Table: data_declaration The data table declaration provides user to specify data set in the format similar to spread sheet. FORMAT DATA_TABLE (inp1, [inp2, ...] out ) inp1_val_0 [,] [inp2_val_0 ...] out_val_0 [,] inp1_val_1 [,] [inp2_val_1 ...] out_val_1 [,] inp1_val_2 [,] [inp2_val_2 ...] out_val_2 [,] . . . ; Where - inp* are input variable names, i.e. must be IN, INOUT, or LOCAL variables - out is the output variable, i.e. must be OUT, INOUT, or LOCAL variable - inp1_val_* are values of variable inp1 - inp2_val_* are values of variable inp2 - out_val_* are values of output out The input variable names and output variable name declared within the parenthsis determine the contents of the data table entries and the order of data values appeared in the table. Note that the out value should be the output based on the input values before it. Normally, this is the case of data table generated from a model. For measured data, the skew of input and output should be considered before building the data table. Configuration: configuration_declaration When starts up, GAF searches configuration file "GAF.CFG" under current directory or where GAF.EXE resides. The configuration file is a text file for setting up the GAF environment. The format of the configuration file is a series of "attribute = attribute_value;". The attribute can be in any order and can be defined multiple times with the latter value overrides the previous value. FORMAT attr_name = attr_val; WHERE - attr_name is the name of the attribute, availble attributes are listed below - attr_val is the valid value for the specified attribute FCL allows configuration information been declared within "CONFIGURATION" and "END_CONFIGURATION;" keywords to override those defined in the gaf.cfg file. The format for declaring configuration information in a segment is: FORMAT CONFIGURATION attr_name = attr_val; . . . END_CONFIGURATION; The following configuration attributes are provided by GAF. COLOR_MODE Enable/disable color display. Valid values are YES and NO. The default is COLOR_MODE = YES; MENU_COLOR Sets the color for the system menu and pulldown menus. Available colors are: BLACK, BLUE, GREEN, CYAN, RED, MAGENTA, BROWN, LIGHTGRAY, DARKGRAY, LIGHTBLUE, LIGHTGREEN, LIGHTCYAN, LIGHTRED, LIGHTMAGENTA, YELLOW, WHITE. The default is MENU_COLOR = CYAN; MENU_TEXT_COLOR Sets the foreground text color for the system menu and pulldown menus. The default is MENU_TEXT_COLOR = BLUE; HIGHLIGHT_COLOR Sets the color for the highlight bar. The default is HIGHLIGHT_COLOR = GREEN; HIGHLIGHT_TEXT_COLOR Sets the foreground text color for the highlighted entry. The default is HIGHLIGHT_TEXT_COLOR = RED; DISPLAY_MODE Initializes the display mode. Available values are: NORMAL_DISPLAY, DISPLAY_PLOT_ONLY, DISPLAY_WHOLE, DISPLAY_NO_PLOT, DISPLAY_TEXT_ONLY. Please refer to Display section for details. The default is DISPLAY_MODE = NORMAL_DISPLAY; WINDOW_SPLIT Initializes the size of the text window. The value is in percentage of the screen. Please refer to Display section for details. The default is WINDOW_SPLIT = 55; PLOT_DURATION Assigns the duration of the trend plot in seconds. The default is PLOT_DURATION = 5.0; TEXT_FONT FONT_SIZE Configures the text font. Available font are: NORMAL and SMALL. The size should be 1 for normal font and 4 for small font. The default is TEXT_FONT = NORMAL; FONT_SIZE = 1; DISPLAY_LOCAL_VARIABLE This attribute determines whether local variables will be displayed along with input/output variables. The default is DISPLAY_LOCAL_VARIABLE = NO; SIM_SCHEDULE_RATE The SIM_SCHEDULE_RATE determines the simulated schedule rate. The value should be a floating number in seconds. The default is SIM_SCHEDULE_RATE = 0.06; SIM_INC_RATE The SIM_INC_RATE attribute defines the increment of + and - keys during simulation mode. The default is SIM_INC_RATE = 0.005; EVAL_SAMPLE This attribute defines how many samples should GAF take for calculating average evaluation result during adaptation. The default is EVAL_SAMPLE = 3; GENE_WEIGHT_INC Defines the proportion weight increment for gene pool. The default is GENE_WEIGHT_INC = 2; MAX_BEST_ITEM Initializes the gene pool size, i.e. number of best genes saved in the best list. The default is MAX_BEST_ITEM = 10; TEST_TIME The TEST_TIME attribute initializes the duration for each evaluation cycle for adaptation. The value is in seconds, and default is TEST_TIME = 2.0; MIN_SCORE This attribute sets the minimum evaluation score for the adaptation mode. The default is MIN_SCORE = 0.6; CHG_RULE_WEIGHT ADJ_INPSET_WEIGHT ADD_RULE_WEIGHT ABLE_RULE_WEIGHT CHG_CYCLE_WEIGHT CHG_GAIN_WEIGHT These attributes are used to assign the weight for genetic adapting methods. The values should be in integer, and the defaults are 1. Please refer Method section for details. ADJ_INPSET_MUTATE_WEIGHT ADJ_INPSET_CROSSOVER_WEIGHT ADJ_INPSET_INTENSIFY_WEIGHT ADJ_INPSET_BROADEN_WEIGHT ADJ_INPSET_SHIFT_WEIGHT ADJ_INPSET_CHG_TRUTH_WEIGHT ADJ_INPSET_SCALE These attributes are used to assign the weight for adjusting the input fuzzy membership set. The values should be in integer, and the defaults are 1. Please refer Method section for details. MAX_RULE The MAX_RULE defines the maximum number of rules allowed during adaptation. The default is zero, i.e. there is no limit for number of rules. MAX_CYCLE_RATE MIN_CYCLE_RATE These two attributes define the limit of cycle rate for the adaptation. The default are: MAX_CYCLE_RATE = 0.50; MIN_CYCLE_TIME = 0.03; EVAL_NULL_BAND Defines the null band for the score of evaluation result. The default is EVAL_NULL_BAND = 0.02; ENABLE_ADAPT_LOG ENABLE_DETAIL_ADAPT_LOG ENABLE_PLOT_DATA_LOG These attributes initialize the status of enable or disable adapt log, adapt detail log, or the plot data log. Valid values are: YES to enable and NO to disable. The defaults are NO. ADAPT_LOG_FILE PLOT_DATA_FILE ADAPT_REPORT_FILE STATISTIC_REPORT_FILE These attributes define the file names for the adapt log, plot data, adapt report, or the statistic report. The defaults are: ADAPT_LOG_FILE = "adaptlog.log"; PLOT_DATA_FILE = "adaptlog.dat"; ADAPT_REPORT_FILE = "adaptlog.rpt"; STATISTIC_REPORT_FILE = "adaptlog.sta";