Accounting for cost heterogeneity on the demand in the context of a technician dispatching problem

In the technician dispatching problem, a given number of repair teams must visit different locations to provide service support. Considering that there is a fixed vehicle capacity and variations in the demand, not all requests can be satisfied on time and therefore some of them must be delayed. Most implementations of the dispatching problem consider a penalty that might vary depending on the customers to internalize that they have heterogeneous costs for being postponed. In this research we analyze how such variations in costs affect the outcome of service planning in the context of an efficient technician dispatching problem. We focus our analysis on two objectives: first, to understand how cost heterogeneity affects the performance of optimal solutions, and second to illustrate how a firm could implement an ad-hoc methodology even in cases where only observable customers’ features can be traced. Specifically, we explore how the distribution of costs affects optimal solutions of allocating teams during a daily operation of the service provider, and then we propose a Markovian model to capture cost-heterogeneity for the case where the cost of failure can be traced to observable operational characteristics. In this model we explicitly consider the cost faced by the customer by having inferior service quality. Our results indicate that when customers are sufficiently different, transportation and total penalty costs decrease gaining in operational efficiency. Moreover, results from the Markovian model indicate that firms can take advantage of these operational gains even in cases where only few customer characteristics are observed.