6.4 Evaluation of Location Alternatives

The simulation model was used to evaluate several alternative fire station location patterns, some of which were derived from the models in Section 5.0 and some of which were suggested by Alexandria officials. This was accomplished by comparing the operating statistics given by specific alternatives with those from the actual current configuration of stations.

These alternatives were:

1. Optimal location of fire stations as given by the Maranzana
location method

2.

3.

Fire station locations as given by the REDIST method

Add an eighth station in the western part of Alexandria

4. Move the engine company in Station 1 into Station 3

5.

Add Station 8; move the engine company currently in Station 1
to Station 3

6. Move the engine company in Station 1 to the new Station 8 location.

For each alternative, statistical distributions were collected separately for cases where one, two, and three engine companies responded, as shown in Table 10.

It is desirable to lower both the mean response times and the standard deviation (which measures the variability of the responses), by reducing the number of long response times. The differences in Table 10 are not sufficiently large to warrant conclusions without statistical analysis. Since the Alexandria Fire Department was actually planning to add an eighth station, further analysis was done on this alternative. Specifically, the response time statistics from the simulation model for this case were compared with those for the present station locations.

The distribution of response times for the last arriving engine produced varying results depending on whether 1, 2, or 3 engines were used for service.

Figures 14, 15, and 16 graphically show the frequency of different response times. The two lines are coded with symbols to indicate the alternative each represents. The two histograms in Figure 14 are close together over the whole response time range, and no appreciable difference is observed. However, Figures 15 and 16 show that there were a number of cases in which the second or third engine company dispatched would arrive sooner if Station 8 were added.

A standard statistical test (Kolmogorov-Smirnov) was used to determine if the two alternatives differed significantly. For one and two engine cases, no significant difference was found. However, However, three engine cases showed a significant difference at the 1% level of significance. Consequently, it may be concluded that the addition of Station 8 would significantly reduce last-engine response time for cases requiring three engine companies.