An efficient two-phase algorithm for cyclic days-off scheduling

Abstract

Workforce scheduling is a very important and practical problem, especially in service industries. The days-off scheduling problem arises in organizations that operate 7 days a week. Since workers must be given weekly breaks, they are assigned to different days-off shifts (work patterns). The objective is to satisfy daily labor demands with a minimum number or cost of workers. This paper considers a specific type of days-off scheduling, in which each work pattern contains five work days and two consecutive off days per week. An algorithm for solving this problem, with minimum labor cost, is proposed and shown to be computationally more efficient than other available methods. An efficient two-phase algorithm is presented for solving the cyclic manpower days-offscheduling problem with two consecutive off days per week. First, a simple formula is used to calculate the minimum workforce size. Next, this minimum is incorporated as a constraint in a continuous linear programming (LP) model of the problem, ensuring an optimum integer solution. The new algorithm can handle the case in which the costs of different days-off patterns are not equal. Computationalexperience with 1250 test problems shows this algorithm to be significantly more efficient than other available solution methods.