Keywords: evolutionary algorithms, optimization, decision making, fire.

Determining high performance strategies for a fire brigade's use and deployment of resources in order to minimise the risk of loss of life and property is a highly complex multi-variable problem. Software is available to fire and rescue services within the United Kingdom which allows them to investigate the effectiveness of different response strategies and deployments of resources, such as fire station location, staffing levels and fire appliance allocations. This software, called the Fire Service Emergency Cover Toolkit (FSEC), is based on a geographical information system and contains the information and procedures needed to define the risks, operational and geographical relationships for a given fire brigade, but only allows the evaluation of one potential solution at a time. It is also graphics-intensive software and has long execution times (of the order of 20 minutes to simulate a typical brigade). Thus, the evaluation of multiple possible solutions is only really practical for a very small number of solutions.

This paper demonstrates that the total number of potential configurations for the resources of a given fire brigade is truly massive (around 1055 for a typical brigade). It is impossible to evaluate all of these manually in order to find configurations which minimise the loss of life and property. The paper outlines the development of novel software to find high performance solutions for the deployment of fire brigade resources, which couples evolutionary algorithms with the methodology of the FSEC toolkit. Evolutionary algorithms allow the relatively rapid identification of areas of good potential solutions by sampling only a small percentage of the total search space. As such, they offer a relatively efficient method of searching for effective resource strategies. In order to achieve this, the methodology used in the FSEC toolkit to predict life and property loss based on a given resource configuration has been re-written in a more computationally efficient model using Fortran 90 resulting in significant reductions in execution time. The model can now evaluate multiple solutions in around 3% of the time taken with the original FSEC software.

The paper also gives details of the linking of the FSEC-based model with evolutionary algorithms in order to find effective solutions to the problem, and concludes that the combination of an efficient fitness function based on FSEC with evolutionary algorithms represents a breakthrough in the ability of the UK Fire and Rescue Services to optimise their use of resources.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £125 +P&P)