Constrained multi-global optimization using a penalty stretched simulated annealing framework

Abstract

This paper presents a new simulated annealing algorithm to solve constrained multi-global optimization problems. To compute all global solutions in a sequential manner, we combine the function stretching technique with the adaptive simulated annealing variant. Constraint-handling is carried out through a nondifferentiable penalty function. To benchmark our penalty stretched simulated annealing algorithm we solve a set of well-known problems. Our preliminary numerical results show that the algorithm is promising.