The impact of neighbourhood size on the process of simulated annealing: Computational experiments on the flowshop scheduling problem

Simulated annealing (SA) has been applied to many combinatorial optimisation problems in a variety of areas of application. In theory, the procedure converges to the global optimal solution provided that the computation time is long enough, which is often unrealistic. In applications, it is essential to get a good solution within a reasonable time. It has been shown that the practical performance of SA depends very much on the selection of parameters. However, parameters related to neighbourhood have not received much research, especially their effect during the whole process of SA. This paper investigates the effect of neighbourhood size on the SA process through computational experiments on the flowshop scheduling problem and proposes an improved SA procedure with a variable neighbourhood size. Experiment results show that the comparative performances of SA with different fixed neighbourhood sizes are related to computation time allowed and that the proposed SA with variable size always gives excellent performance especially on the effectiveness of the whole process.