Service quality, employee satisfaction, and cross-training: A broader look at workforce scheduling and rostering

Service level, the percentage of customers served within a predetermined time interval, is the most widely used quality indicator in the service industry in personnel scheduling; however, it is doubtful that service level alone represents as broad a measure of service quality as management would desire. The dubious relationship between service quality and service level is what inspired this study. With modern trends of providing more complicated services for more sophisticated systems in a more competitive market, there are many criteria other than service level that need to be considered in order to make good workforce scheduling and rostering decisions. This problem is especially difficult when dealing with systems having time-varying stochastic demand throughout the day, namely M(t) /G/s(t) queueing systems, which are widespread in service industry.; The traditional scheduling and rostering approaches require the application of analytical queueing and integer linear programming methods. Though currently well accepted in the industry, these methods have several limitations. Analvtical models involve highly simplified queueing systems and/or nonrealistic assumptions, and might not be able to generate all service quality indicators management is interested in. Obtaining optimized schedules and rosters for a large system is very tedious, and optimality is not always guaranteed. Scheduling and rostering are typically conducted as two independent processes; consequently, employee satisfaction is completely ignored during scheduling and two optimization procedures are needed. In addition, traditional approaches lack the ability to consider multiple service quality and employee satisfaction criteria simultaneously and cannot handle employees who are cross-trained in multiple tasks.; A novel multi-objective framework which overcomes these many limitations is proposed in this study. Both schedules and rosters are generated in one integrated step; among all efficient ones recognized the best schedule and/or roster is chosen with the direct input of management, in accordance with their particular business strategy. This framework can also potentially be used for cross-trained employees. Cross-training strategies are identified for systems with various demand curve patterns and levels in terms of employee pooling and timing. These strategies will provide guidelines to expand the proposed approach in a multi-tasking setting in the future.