SwarmOps Differential Evolution (DE)
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 DE is a constrained optimization that uses a swarm of agents (e.g., candidate designs). At each iteration, the agents are moved around the best known design and the swarm is expected to converge near to a best design. This method handles only continuous design variables.
More Information
The algorithm starts by randomly generating a swarm of size np. Trial points for next generation are generated by a combination of the known best design, current design, and other designs. Pick a random integer value R between 1 and the dimension of the search space (n). The next trial point is determined by using the following formula.
where yi is the i-th coordinate of the new trial point, xi is i-th coordinate of the current design, and gi is i-th coordinate of the known best design so far. F is the differential weight, and CR is the crossover probability. If the new trial point improves the current design, the current design is moved to the trial point. Otherwise, it stays the same. As the iteration goes on, the swarm is expected to gather around a best design.
References
SwarmOps Manual (p. 17)
Control Parameters
| Name | Default Value | Description |
|---|---|---|
| Optimization Parameters | ||
| CR | 0.496 | Probability of crossover being performed. Thus the value specified must be between 0 and 1. |
| F | 0.5313 | Differential weight. Option must have a value greater than 0. |
| NP | 37 | 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. | |
| 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. |