Dynamic Repositioning of Indistinguishable Service Units on Transportation Networks

Abstract

The repositioning of urban emergency units is examined analytically using the methodology of Markovian Decision Processes. In practice urban emergency vehicles (e.g., fire engines) are constantly subject to repositioning in order to assure a proper posture for responding quickly to future demands. The analysis assumes that there are q indistinguishable servers on the network located initially at q nodes—“The Home Locations.” Depending on the status of other server locations, any available server can be moved to any other vacant location in the network. The states of the system are defined according to the status of each location (occupied or vacant). The policy space consists of decisions on where and when to move service units for any possible state. The objective is to minimize the long term expected cost (in time units) of operating the system.