A Dual Sequence Simulated Annealing Algorithm for Constrained Optimization

Abstract

We propose a dual sequence Simulated Annealing algorithm, DSAC, for solving constrained optimization problems. This approach eliminates the need for imposing penalties in the objective function by tracing feasible and infeasible solution sequences independently. We compare DSAC with a similar single sequence algorithm, PSAC, where the objective function is augmented by various penalty forms related to constraint infeasibilities. Numerical experiments show that DSAC is more effective than its counterpart PSAC in the worst and average case performances.