State-transition simulated annealing algorithm for constrained and unconstrained multi-objective optimization problems

Abstract

In this article, a novel multi-objective optimization algorithm based on a state-transition simulated annealing algorithm (MOSTASA) is proposed, in which four state-transition operators for generating candidate solutions and the Pareto optimal solution is obtained by combining it with the concept of Pareto dominance and then storing it in a Pareto archive. To ensure the uniform distribution of the Pareto optimal solution, we define a crowded comparison operator to update the Pareto archive. Simulation experiments were conducted on several standard constrained and unconstrained multi-objective problems, in which convergence and spacing metrics were used to assess the performance of the MOSTASA. The test results manifest that the MOSTASA can converge to the true Pareto-optimal front, and the solution distribution is uniform. Compared to the performance of other multi-objective optimization algorithms, the proposed algorithm is more efficient and reliable.