Performance evaluation of acceptance probability functions for multi-objective SA

A probabilistic local search algorithm called simulated annealing (SA) is a useful approximate solution technique for multi-objective optimization problems. When we use SA to solve multi-objective optimization problems, we cannot use an acceptance probability function used for single objective optimization problems. Therefore, several types of acceptance probability functions for multi-objective SA have been previously proposed. In this paper, we introduce a parameterized acceptance probability function for multi-objective SA, which changes its type depending on the parameter, and investigate how the performance of the multi-objective SA depends on the type of acceptance probability function in two test problems.