Memetic Algorithm for the Nurse Scheduling Problem

Abstract

The Nurse Scheduling Problem (NSP), like the well-known Travelling Salesman Problem (TSP), is an NP-hard problem. In this study, we use a tailor-made meta-heuristic Memetic Algorithm (MA) to optimize the NSP. The MAis a hybrid algorithm, being a combination of the Genetic Algorithm (GA) and a local search algorithm. The performance of the MA is found to be superior to that of a solitary algorithm like GA. The MA solves the NSP in two stages. In the first stage, the randomly generated solutions are evolved till they become feasible (i.e., the hard constraints are satisfied) and in the second stage, these solutions are further evolved so as to minimize the violations of the soft constraints. In the final stage, the MA produces optimal solutions in which the hard as well as the soft constraints are completely satisfied.