A Patrol Car Allocation Model: Background

Before designing a computer program for allocating police patrol cars by time and geography, a review was undertaken of previously existing programs of this type. Nearly all of the programs calculated queuing statistics for the collection of patrol cars by assuming a steady-state system with calls for service arriving within priority levels according to Poisson processes and having independent, identical, exponentially distributed service times. Unavailabilities of patrol cars for reasons other than calls for service were handled in the models either by artificially increasing the arrival rate of calls or by assuming that the number of servers is smaller than the number of patrol cars. Some programs calculated additional performance measures such as travel times and preventive patrol frequencies. All the programs had the capabilities to describe performance statistics for an allocation proposed by the user, but they differed in their capabilities to prescribe desirable allocations. None of the programs had achieved general acceptance because each had virtues and inadequacies not present in the others.

• 저널Management Science · 24(12) · 1280–1290
• 토픽Facility Location and Emergency Management · Organizational Behavior and Human Resource Management
• DOI10.1287/mnsc.24.12.1280
• 저자Jan M. Chaiken, Peter Dormont