A decomposition heuristic for short-term planning of assessment centres

Abstract

In an assessment centre, several candidates for a job vacancy perform a set of predefined tasks while being observed and evaluated by so-called assessors. For the organisers of such assessment centres, a challenging job is to schedule the tasks and assign the prescribed number of assessors to the tasks such that the total waiting time for the assessors is minimised. This planning situation has been reported to us by a human resource management service provider. Application-specific restrictions distinguish this problem from related scheduling problems discussed in the literature, e.g., the resource-constrained project scheduling problem. We present a mixed-integer programming-based decomposition heuristic, which iterates between pre-scheduling, assignment, and re-scheduling subproblems. Our computational results demonstrate that this novel heuristic outperforms the state-of-the-art approaches on a set of 240 benchmark instances. Furthermore, this heuristic provides optimal solutions to a set of four real-life instances. [Received 30 September 2016; Revised 26 January 2017; Accepted 27 April 2017]