Study Of Contingency Plan Generating Technology For City Fire Fighting

Nowadays, the fires in city become more and more complex and dangerous, andthe elements involved are much more various. It is urgent to quickly make an accurate,reasonable and comprehensive fire contingency plan, which can guide the rescueoperations and reduce casualties and property losses. Therefore, using computertechnology to assist the generation of city fire contingency plan has become a currentresearch focus and application trends.However, there are still some obvious deficiencies of applying existing researchresults in practice. These deficiencies are mainly manifested in the following aspects.Firstly, the generation of fire contingency plan only relies on a rule-based reasoningframework or case-based reasoning framework. Secondly, they have not considered theweights of each attribute on the cases or the weights set are not reasonable during thecase retrieval. Thirdly, the distribution of fire rescue forces is limited to localoptimization, failed to achieve the global optimal solution. To conquer the problemsmentioned above, the paper introduces a method of combining the case-based andrule-based reasoning framework to generate a contingency plan; and provides a methodof case retrieval based on the characteristic weight auto-adapted algorithm to improvethe reference of the recommended case; then puts forward a distribution strategy of firerescue forces based on the genetic algorithm to achieve the overall-optimization in theprocess of force distribution. Specifically, the main contributions of this thesis include:(1) In view of the shortcomings when separately using the case-based reasoning(CBR) and rule-based reasoning (RBR) framework, this paper synthesize these twoframes together to generate a new fire contingency plan. First of all, the new methoduses the CBR framework to generate the plan, if there is no case recommended or thecase recommended is not suitable enough, and then the RBR framework will be served.(2) By modeling the process of case retrieval, this paper proposes a characteristicweight auto-adapted algorithm, which determines the weight values of case according tothe coverage of each attribute in the case library and computes the overall-similarity bythe weighted sum of local-similarity. By this way, the more reliable case is beenrecommended to the user. The validity of the new method has been validated by theexperiments.(3) Aimed at the simultaneous rescue demand of several city fires, this paperproposes a strategy of force distribution based on the genetic algorithm and designs thescheme of gene coding to solve the problem of force distribution. The feasibility of theforce distribution has been demonstrated by the experiments.(4) Under the background of city fire emergency rescue, taking the generation ofemergency rescue plan as the purpose, this paper designs a prototype system for thegenerating of emergency rescue plan. The city fire information collection, case retrievalbased on CBR, emergency rescue plan generation based on RBR and the auxiliary decision effect brought by the geographic space information are showed in the prototypesystem.