home *** CD-ROM | disk | FTP | other *** search
- Introduction to GAF
- GAF (Genetic-Adapt Fuzzy Control) runs in IBM PC and compatible machines.
- The development of GAF is intended for off-line simulation, adaptation
- and on-line adaptive control. Currently the functions for simulation and
- off-line adaptation are completed, the on-line portion is still under
- developing.
-
- GAF builds run time data directly from fuzzy IF-THEN rule segments
- contained in simple text files. It eliminates intermediate compilation and
- achieves quick turn-around time.
-
- A segment, a basic unit in GAF, can be defined by means of:
- fuzzy rules
- mathematic formulas
- combination of fuzzy rules and math formulas
- Not only the control can be defined in segments, the feedback and the
- evaluation can also be defined for simulation and adaptation. GAF also
- supports the use of data set (measured or modeled) as its feedback.
-
- With GAF's graphic display and user interface it also provides user as an
- education tool for understanding fuzzy logic and genetic algorithm.
-
- GAF allows users to generate a fuzzy control system by simply defining
- the inputs, outputs, data set, and initial rule sets. GAF uses genetic
- algorithm to derive proper rules and fuzzy sets from the initial rules.
- By changing, adding, deleting rules and fuzzy membership sets, the genetic
- algorithm automatically adapts and optimizes the fuzzy control system.
-
- GAF's Simulation
- GAF provides an integrated simulation environment for user to fine-tune
- their fuzzy control applications and examine the response of the fuzzy
- rules with certain conditions (i.e. for some known input values). High
- lights of GAF's simulation environment are:
-
- Integrated simulation
- Direct use data set
- Verify single segment
- Change schedule rate
- Single step to view details of fuzzy inference
- Change output gain
- Enable/disable rule
-
- GAF's Off-line Adaptation
- With evaluation segment, GAF is capable of adapting user's fuzzy control
- applications automatically. In future release GAF will provide user with
- canned evaluation method. Major functions are:
-
- Change existing rule
- Change fuzzy membership set
- Adding new rule
- Disable existing rule
- Alter cycle time
- Alter output gain
-