The GA Procedure
- Overview
-
Getting Started
-
Syntax
PROC GA Statement ContinueFor Call Cross Call Dynamicarray Call EvaluateLC Call GetDimensions Call GetObjValues Call GetSolutions Call Initialize Call MarkPareto Call Mutate Call Objective Function PackBits Call Programming Statements ReadChild Call ReadCompare Call ReadMember Call ReadParent Call ReEvaluate Call SetBounds Call SetCompareRoutine Call SetCross Call SetCrossProb Call SetCrossRoutine Call SetElite Call SetEncoding Call SetFinalize Call SetMut Call SetMutProb Call SetMutRoutine Call SetObj Call SetObjFunc Call SetProperty Call SetSel Call SetUpdateRoutine Call ShellSort Call Shuffle Call UnpackBits Function UpdateSolutions Call WriteChild Call WriteMember Call
-
Details
Using Multisegment Encoding Using Standard Genetic Operators and Objective Functions Defining a User Fitness Comparison Routine Defining User Genetic Operators Defining a User Update Routine Defining an Objective Function Defining a User Initialization Routine Specifying the Selection Strategy Incorporating Heuristics and Local Optimizations Handling Constraints Optimizing Multiple Objectives
-
Examples
- References
| Incorporating Heuristics and Local Optimizations |
It is often effective to combine the genetic algorithm technique and other local optimizations or heuristic improvements. This can be done within the GA procedure by incorporating a local optimization into a user objective function and returning an improved objective value. Either your user objective function can replace the original solution with the optimized one, or you can leave the solution unchanged, replacing it with the optimized one only at the final iteration.
Replacing the original solution with the locally optimized one speeds convergence, but it also increases the risk of converging prematurely. If you choose to do so, you can modify the solution by writing the changed solution back to the population with a WriteMember call. You could also consider replacing the original solution with some probability p. For some problems, values of p from 0.05 to 0.15 have been shown to significantly improve convergence while avoiding premature convergence to a local optimum. This technique is illustrated in Example 3.1.