This paper describes a computer program that uses simple rules and applies them probabilistically to minimize manpower requirements in shops of a naval shipyard. Preliminary runs on a small fictitious problem show that this “multiship, multishop, workload-smoothing” program, called (MS)2, may be able to reduce the maximum number of men required on any one day by 30 to 60 percent.
The program employs the critical path method of analyzing the jobs to be done on a given project (ship) and begins by assigning all jobs at their earliest possible start time. Then suitable jobs on peak workload days are selected at random and shifted to later days beyond the peak days until further shifting no longer reduces peak loads. The first segment of the program smooths the workloads on all shops simultaneously. After the first segment is completed the second segment performs further smoothing on individual shops, beginning with the most expensive shop.
Although a limited amount of experience has been gained with this program, it appears likely that it can be developed further so that it can handle realistic-sized problems such as those encountered in a Naval Shipyard. Moreover, it can be applied on three levels: (1) smoothing shop workloads in a shipyard; (2) smoothing the workloads of individual shops; and (3) smoothing shipyard workloads.