An investigation of initial solutions on the performance of an iterated local search algorithm for the permutation flowshop

This paper examines the effect of initial solutions on the performance of an iterated local search (ILS) algorithm for the permutation flowshop problem with the objective of minimizing total flowtime. An ILS algorithm is applied to a set of test problems, and in each separate trial the algorithm is started from an initial solution generated by one of six different methods. Experimental results indicate that initial solutions generated by a neural network are more effective in promoting the performance of the ILS algorithm towards better solutions. A modified version of the ILS algorithm, in which an initially restricted neighborhood search is gradually expanded with each iteration, is also proposed and tested. The results from this modified ILS compare very favorably with published results from a traditional ILS approach.