ALGORITHMIC SUPPORT OF THE INTELLECTUAL DECISION SUPPORT SYSTEM FOR THE HEAD OF FIREFIGHTING

During firefighting real fires the heads of firefighting often take insufficient effective solutions: ineffective solutions are up to 57 % of the total number of decisions. The rational way to increase the efficiency of head decision is the development and application of specialized software — the Intelligent Decision Support System. The authors have proposed to design and use the Intelligent Decision Support System based on the Adaptive Neural Fuzzy Inference Systems (ANFIS) for solving the problem of forecasting the dynamics of the spread of fire. The authors have developed the mathematical model based on network ANFIS: the mathematical model for forecasting of fire area; the mathematical model for forecasting of firefighting area; the mathematical model for forecasting of the thermal radiation density in the room; the mathematical model for forecasting of the neutral zone height in the room. The authors have carried out testing of this system and have identified: 1) the network for forecasting of fire area — the average testing error on the training data was 13.97 m2 , and on the test sample was 15.3 m2 ; 2) the network for forecasting of firefighting area — the average testing error on the training data was 15.4 m2 , and on the test sample was 16.7 m2 ; 3) the network for forecasting of the thermal radiation density in the room — the average testing error on the training data was 0.07 kW
m2 , and on the test sample was 0.15 kW
m2 ; 4) the network for forecasting of the neutral zone height in the room — the average testing error on the training data was 0.023 m, and on the test sample was 0.031 m. As the final stage of development of the Intelligent Decision Support System we obtained the certificate of state registration of software. The expected cost-effectiveness of the implementation of the Intelligent Decision Support System for the most likely fire in the port area “Kaliningrad Sea Commercial Port” is 7028300 rubles. The scientific novelty of the research is that the first time: 1) the authors have developed methods and models of decision making of head of the firefighting under uncertainty and lack of time; 2) the authors have developed the mathematical model for forecasting of fire in buildings, which consists of the following mathematical models:
for forecasting of fire area;
for forecasting of firefighting area;
for forecasting of the thermal radiation density in the room;
for forecasting of the neutral zone height in the room; 3) the authors justified the use of an effective means of artificial intelligence — the Intelligent Decision Support System to solve complex multicriteria problem of forecasting of fire in buildings in conditions of uncertainty; 4) authors have created new Intelligent Decision Support System for the head of firefighting in the seaport.