Research Of Decision Support Algorithms Based On Information Fusion Algorithm Management System

Multi-Sensor Information Fusion (MSIF) is the process that analysis, processing, integrated from multiple sensors or sources, resulting in the synthesis of new and more effective information. MSIF works through the combination of different algorithms. Different algorithms vary significantly in fusion performance according certain maneuvering target. The research emphasis in the paper is the decision support methods of information fusion management system. The system can get optimal algorithms combination by feedback according the motion character of different maneuvering target, to mark algorithms selected automatically and help to make a decision.Firstly, analysis traditional decision support methods and find out their disadvantages. A new multi-objective decision method different from other methods was adopted, and make a research based on the information fusion management system. The main algorithm is the Fast Non-dominated Sorting Genetic Algorithm based on Pareto (NSGA-Ⅱ). The algorithm has the ability about the Approximation of Pareto front because of the fast non-dominated sorting mechanism. At the same time, diversity of Pareto optimum solution can be preserved though the Sort operation according the crowding distance among different individuals. Then, considering the disadvantage about NSGA-II when generating sub-populations, an improved method with distribution function was raised, and the diversity of sub-population can be preserved through new elitist strategy. According the simulation experiment, the new method has a better effect on the diversity and convergence, and can get the better algorithms combination under the same condition.