Heuristic Algorithm for Workforce Scheduling Problems

In this paper we present a heuristic approach for solving workforce scheduling problems. The primary goal is to minimize the number of required workers given a pre-established shift demand over a planning horizon. The proposed algorithm starts with an initial solution (initial number of workers and their shift assignment) and iteratively searches the state space, moving towards better solutions via a local search procedure. Local optima are avoided by guaranteeing that the algorithm never returns to a previously visited solution. The algorithm stops after a termination criterion is met. The solution provides a detailed schedule of each worker on each shift. A number of constraints such as minimum and maximum number of working hours, rest days, and maximum number of continuous working hours are considered. The algorithm was tested on a number of randomly generated problems of different sizes. A Mixed Integer Programming (MIP) formulation is proposed and used as a benchmark. Computational experiments show that the algorithm always found optimal or near-optimal solutions with signifi cantly less computer effort.