SwarmOps Local Unimodal Sampling (LUS)
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 LUS is a constrained optimization algorithm that uses random
sampling around the current design. It may handle noisy objective
functions better than gradient based methods. But the convergence
is expected to be slow due to its random sampling technique.
More Information
It starts with a randomly selected initial design. Uniform random sampling
is used to generate next trial point. Initially the bounds for random
sampling include the entire design space. If the new trial point improves
the current design, it becomes the new current design. If the new
trial point does not improve the current design, the sampling range
is reduced by a factor, q, and another design is tried. The
reduction factor is computed using a formula below.
where n is the number of dimensions.
References
SwarmOps Manual (p. 16)
Control Parameters
| Name | Default Value | Description |
|---|---|---|
| Optimization Parameters | ||
| Gamma | 3 | Gamma parameter to control the search space reduction factor. Option must have a value greater than 0.5. |
| 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 convergence criteria must be met to indicate convergence. Thus option must have a positive integer value. |
| MaxFunctionEvaluations | 2000 | Maximum number of iterations. Thus option must have a positive integer value. |