Tour scheduling with dynamic service rates

This paper examines the benefit of incorporating a group of employees that exhibit dynamic service rates into scheduling tours in a service operation. The service operation that is examined includes a fully productive core (full-time) workforce along with a contingent (full-and part-time) workforce that experiences the learning effect. Two methods that account for the learning effect are analyzed along with two methods that do not consider learning effects. The schedules generated by each method are tested in a simulation of the service environment. The results of a full-factorial experiment indicate that methods that account for learning effects will yield superior solutions over a variety of operating conditions when compared to alternative methods that do not consider learning effects. The performance improvement of schedules generated with the most precise learning curve method was substantially and significantly better than the other methods. The conditions in which the learning curve methods provide the most benefit are explored.