213 COST VARIANCE INVESTIGATION WITH FUZZY SETS T.W. Lin and D.E. O'Leary l 1 Both of the School of Accounting, university of Southern california, Los Angeles, California, 90089-1421 (USA) 1 INTRODUCTION Cost variance investigation is critical to process control. Cost variances provide a report that indicates the deviation between actual costs and standard costs for process systems. Since standard costs are a priori estimates, it is easy to understand that actual and standard costs are not likely to be identical. Thus, the problem that faces managers who receive variance reports on these systems, is to determine which reported cost variances indicate that the system they represent should be examined further. If a system is to receive further examination then a team of investigators will take samples from the actual behavior of the system to estimate if the system is in-control or out-of-control. Thus, managers must determine the extent of the work that is done to determine the state of the system. Ideally, in-control systems are not investigated and only out-of-control systems are investigated. However, in-control systems may be investigated and out-of-control systems may not be investigated because the variance report is only a probabilistic evaluation of the state of the system. If the system is in-control and the system is investigated then a cost is incurred when it is not necessary to incurr a cost. However, if an out-of-control system is not investigated then the system will continue to be out-of-control and the organization will incur the costs of the system being out-of-control. Since there are only limited time and monetary resources, and interrupting a process to investigate has a cost, management's team cannot investigate every system with a reported deviation or variance. Alternatively, if the manager does not decide to investigate the system based on the reported variances then the