A Simple, Adaptive Bubble Search for Improving Heuristic Solutions of the Permutation Flow Shop Scheduling Problem

In this paper we study the application of Bubble Search, an extension of priority-based greedy heuristics proposed by Lesh and Mitzenmacher (2006), as an improvement procedure for heuristic solutions of the permutation flow shop scheduling problem. We compare the performance of Bubble Search on different time scales and for different parameter settings. In particular, we demonstrate the optimal parameter setting depends on the size of the instance, and propose a simple adaptive extension of Bubble Search. Computational experiments demonstrate that the adaptive version outperforms fixed parameter settings. They further show that adaptive Bubble Search is competitive with the most effective constructive heuristics in a comparable time scale, and therefore is a promising technique for flow shop scheduling and related problems.