NCJ Number
70381
Journal
Journal of Contemporary Business Volume: 8 Issue: 3 Dated: (1979) Pages: 35-49
Date Published
1979
Length
15 pages
Annotation
A Seattle program which utilized statistical forecasting and computer simulation to budget police patrolling is described.
Abstract
The Patrol Car Allocation Model (PCAM), developed in 1974, is an analytical model which computes various patrol operating characteristics from call for service rates and patrol car staffing levels. In order to forecast the citywide number of calls for police services that are received by telephone operators in the department's communications center, both a traditional trend projection procedure with seasonal adjustment and the new Box-Jenkins method were used. The historical data which served as a basis for the forecasts covered the 65 4-week periods from 1970 through 1974. The forecasted data were for 26 four-week periods for 1975 and 1976. The PCAM program computes the following average patrol characteristics: travel time for patrol cars in minutes; probability that an incoming call for service would be placed in queue; delay (travel time and queue delay); patrol hours per suppressible crime; patrol passing frequency for a random point on the street per hour; and patrol passing frequency times suppressible crime per hour. The city was divided into small patrol areas with each area normally containing 3 to 10 car beats. The model provides patrol characteristics for each watch and each patrol area, as well as aggregate figures for the entire city. The program can be run in two modes: descriptive and prescriptive. Under the descriptive mode, the user enters the number of cars assigned to each watch and area into the program. The results of a computer run show the average patrol characteristics for the car configuration entered. Under the prescriptive mode, the user enters the total number of patrol car hours available city-wide for each 24-hour period. The program then assigns the optimum number of cars to each watch and area. The trend-seasonal and the Box-Jenkins forecasting methods are described in detail. Footnotes with references are provided.