Assessment of the “Evaluation” function in the simulated evolution algorithm

Simulated Evolution is a stochastic evolutionary search strategy. The algorithm repeatedly executes evaluation, selection, and allocation steps in a sequence, until certain stopping conditions are met. The evaluation step assesses the quality of each individual element of a solution with respect to a single or multiple attributes. The outcome of this step, combined with the results of the selection step, play an important role during the allocation step, where the existing solution is perturbed to generate a new solution. This paper attempts to study the effect of the evaluation step on the quality of the solutions produced by the allocation step. Results suggest that, in general, the Unified And-Or (UAO) operator based evaluation scheme performs significantly well compared to other evaluation approaches.