We consider the following version of the auditing problem. A set of jobs must be processed by auditors A1,...,Am. Each job consists of several tasks and there may be precedence constraints between these tasks. There is a due date associated with each job. Each auditor is available during disjoint time periods. Furthermore, s/he has a minimal and maximal working time. If task i is assigned to an auditor Aj, the processing time is pij and the processing costs are cij. A task assigned to auditor Aj can be preempted only at the end of one of the working periods of Aj. In this case it must be continued at the beginning of the next period. One has to assign the tasks to the auditors and find a feasible schedule for the assigned tasks for each auditor such that the sum of the assignment costs and a weighted sum of tardiness values is minimized. A tabu search procedure for this problem is described and computational results are presented. Copyright © 1999 John Wiley & Sons, Ltd.

We consider the following version of the auditing problem. A set of jobs must be processed by auditors A , . . . , A K . Each job consists of several tasks and there may be precedence constraints between these tasks. There is a due date associated with each job. Each auditor is available during disjoint time periods. Furthermore, s/he has a minimal and maximal working time. If task i is assigned to an auditor A H , the processing time is p GH and the processing costs are c GH . A task assigned to auditor A H can be preempted only at the end of one of the working periods of A H . In this case it must be continued at the beginning of the next period.One has to assign the tasks to the auditors and "nd a feasible schedule for the assigned tasks for each auditor such that the sum of the assignment costs and a weighted sum of tardiness values is minimized.A tabu search procedure for this problem is described and computational results are presented.KEY WORDS: audit scheduling; tabu search; time windows 4. the precedence constraints are not violated, and 5. tasks of J J do not start before r J .Given a schedule (for all auditors), the "nishing time C J of J J is the latest "nishing time of all tasks of J J . Our objective is to assign each task i to an auditor A ? G and to construct a feasible schedule such that the sum of mismatching costs and weighted sum of tardiness ¹ J "max +0, C J !d J ,, i.e.