Many applications exist in which a vehicle or set of vehicles have the resp
onsibility of patrolling and providing coverage over a wide area or network
for the purpose of responding quickly to service requests generated random
ly at demand points in the network. One of the decisions in the operation o
f such patrol vehicles is that of determining the home location or circulat
ory loop for free or idle vehicles. A vehicle is said to be idle if it is n
ot responding or attending to a service request. Such home location can be
a point, a set of points in a network, a circulatory loop, or a set of circ
ulatory loops in the network. The actual number of home locations is system
, fleet size, and analyst dependent. Whether a point or loop strategy is im
plemented, the specification of the home locations or loops is usually done
with the objective of improving the response time of idle vehicles to serv
ice requests originating from demand stations. In cases where circulatory l
oop strategy is adopted, the problem is one of determining the best overall
circulatory loop or loops that would improve the overall system response t
ime. In a typical conventional or generalized network, several loops may ex
ist. Unless a systematic procedure is developed, non-optimal loops could be
selected. In this paper, methodologies for determining the best circulator
y loops for patrol vehicles for a given problem scenario in both unidirecti
onal and bidirectional networks are presented. The problem is modelled as a
mathematical programming problem. Example problems are given to illustrate
the application of the approach. The criterion for optimization is the min
imization of the maximum system response time.