INTEGRATING REMOTE SENSING WITH GIS FOR URBAN FIRE RISK MODELLING: A MULTI-AGENT SYSTEM FOR DYNAMIC RESOURCE REDISTRIBUTION *©Sunil BHASKARAN, *Bruce FORSTER, **Nandan PARAMESH, ***Trevor NEAL

Limited dynamic resources (fire fighting staff and fire equipment) have to be allocated on the basis of spatio-temporal factors, which may influence fire hazard and vulnerability (risk) in urban environments. On the one hand there is a need to develop a model that inputs spatio-temporal variables and enables spatial analysis of such variables for the assessment of composite fire hazard and vulnerability in a given place. On the other hand a major issue is to use such a model to develop algorithms, which will assist in the redistribution of available dynamic resources in order to accommodate increasing hazard levels (Bhaskaran et al, 2001). Such a model will not only enable the comparison of existing hazards, but may also provide a basis for the allocation of dynamic resources. This paper examines the active use of such a risk assessment model for deriving a more balanced distribution of dynamic resources employing a multiagent problem-solving model. By using mutlisensor remote sensing data and integrating such datasets with the available cartographic data in a GIS environment, a methodology to develop a risk assessment model is demonstrated in near real time. By a system of assigning influences and biases the risk model may be modified to suit any geographical region. In this dynamic situation, agents in the proposed multiagent system respond to changes in the world by redistributing the resources after engaging in mutual negotiation. It is hoped that this model will lead to a more balanced utilization of resources thus minimizing the overall global risk in the world.