A Late Acceptance Algorithm for the Lock Scheduling Problem

In this paper we present a number of (meta-)heuristics for tackling the lock scheduling problem. The lock scheduling problem is introduced as the problem of minimising both the water usage of the lock and the waiting time of all the ships. The performance of several local search heuristics with the best improving criterion and the late acceptance criterion is examined. First parameter values for the late acceptance list length are determined.With these values all heuristics and criteria are applied to a large test set. From these experiments we conclude that the performance of the heuristics is increased by the late acceptance criterion.