In those areas where travel time is a significant component in servicing a call, urban based models do not sufficiently represent reality. SWAP explicitly incorporates travel time by representing the patrol system in terms of a Markov process model, using a travel state for each region. When applied to Washtenaw County, primary measures of interest were the fraction of time a unit is on patrol in each township and the mean response times to emergency and routine calls. Results indicate that two patrol cars substantially reduced average response times to all types of calls, when compared to using a single car. A third car in one of the 12 townships substantially reduced average response time in that region. A fourth car led to further reductions in all mean response times. The model successively computes the probability that each car is busy and modifies call rates to account for calls that cannot be handled by cars from their region of origin. The model iterates until these probabilities converge. This approach avoids solving a combinationally large problem for multiple cars and has performed efficiently in practice. Tables, appendixes, and 18 references are supplied.