The multi-customer local delivery problem and the siting of repair parts inventories.

Abstract

A local delivery model was developed for a repair facility- stock point system, given one or more supported production lines and each component repaired may require more than one part. Both deterministic and random demands were considered. The objective function was total expected transportation and delay costs per day. In the deterministic case the total cost curve was discontinuous and the optimal delivery policy could only be determined by exhaustive enumeration. A computer simulation model was needed for the random demand case. The simulation model was also extended to allow random issue processing time and a remote warehouse sited close to the repair facility. The results of the simulation showed that point of entry effectiveness and non-local response times were key factors of expected delay costs and that these costs could be reduced through the use of a remote warehouse. More importantly, providing the best support to customers requiring the fewest parts per component repaired will give the minimum expected delay cost.