SwarmOps Self-Adaptive Differential Evolution (JDE)
SwarmOps algorithms are being removed and may not be compatible with future releases of Analyzer. Users should be cautious when using them in models
Note: SwarmOps algorithms are being removed and may not be compatible with future releases of Analyzer. Users should be cautious when using them in models
Version
SwarmOps 3.0
Description
SwarmOps JDE is a constrained optimizer that handles continuous design variables. It is a variation of SwarmOps DE that automatically adjust control parameters during iterations.
More Information
This is a variation of SwarmOps that introduces a scheme to automatically adjust the differential weight (F) and the crossover probability (CR) during optimization. Initial values of the parameters, Finit and CRinit, are specified by users. During an iteration, F is kept the same or a new value is computed from a uniform random distributionU(Fl, Fu). The chance that a new value is computed is given by a probability, τF. Same technique is used to get new CR, where the value is kept the same or it is updated by a uniform random U(CRl, CRu). The chance that a new value is computed is given by a probability, τCR. The two parameters are used to generate a position of a design. If the new position improves the current design, the current design is updated or the current design stays the same.
References
Tuning & Simplifying Heuristical Optimization, Pedersen, E. H. P., PhD thesis, University of Southampton, 2010, pp. 70.
Control Parameters
| Name | Default Value | Description |
|---|---|---|
| Optimization Parameters | ||
| NP | 50 | Number of agents or population size. Option must have a positive integer value. |
| Seed | Random number generator seed value (optional). Specifying the same seed value between two different optimization runs would help generate identical results provided all other parameters stay the same. This is a good way to analyze the effect of different parameters on the optimization. | |
| Optimization Parameters for Crossover | ||
| CRInit | 0.9 | Initial value of the crossover probability. Thus the option must have a value between 0 and 1. |
| CRl | 0 | Lower bound of the crossover probability. Thus the option must have a value between 0 and 1. |
| Crossover | Best | Crossover scheme.
|
| CRu | 1 | Upper bound of the crossover probability. Thus the option must have a value between 0 and 1. |
| TauCR | 0.1 | Probability of changing the crossover probability value. Thus the option must have a value between 0 and 1. |
| Optimization Parameters for Differential Weight | ||
| FInit | 0.5 | Initial value of the differential weight. Option must have a value greater than 0. |
| Fl | 0.1 | Lower bound of the differential weight. Option must have a value greater than 0. |
| Fu | 0.9 | Upper bound of the differential weight. Option must have a value greater than 0. |
| TauF | 0.1 | Probability of changing the differential weight. Thus the option must have a value between 0 and 1. |
| Stopping Criteria | ||
| AbsoluteConvergenceTolerance | 1E-05 | Maximum absolute change in fitness value between successive evaluations to indicate convergence. The value specified must be greater than 0. |
| ConsecutiveFunctionEvaluations | 200 | The number of consecutive iterations for which the absolute or relative convergence criteria must be met to indicate convergence. Thus option must have a positive integer value. |
| MaxFunctionEvaluations | 1000 | The maximum number of iterations. Thus option must have a positive integer value. |