Solving Multi-agent-based Distributed Model Portfolio. Of Idss

Supported by the National Plan foundation—the research on thetechnology of military programming intelligent decision support system, the focus of this paper is on the model system of IDSS. Based on the technologies of model base in intelligent decision support system (IDSS), with the theory of system engineering, standardization of model and object-oriented method, the paper shows a model base in IDSS.IDSS can provide management personnel and decision-makers with necessary information, helping make major decisions with more feasibility. As an effort to improve the efficiency and ability of decision-making and to survive and thrive in intense competitions, many enterprises have spent considerable time and energy on IDSS development. With development of information technology, war in future will become the war of information. It is too hard for the commander to make decision quickly and correctly. At this time, IDSS will play an important roll, helping commanders make perfect decisions quickly.The paper focus on four aspects:First, the paper pays more attention to model base, basing analyzing on structure of IDSS. IDSS is based on model base, which is not a standard system. The paper focuses its work on designing a model base, which is designed by using the theory of object-oriented method.Second, Model combining is implemented. It combines unit models with logic rule to solve the problems. Therefore, it can help make decision easier, solve problems on the distributed network, reuse models and save space. The paper solves several problems on combing model and put forward two methods. One model is on parameter driving, the other is on classifying driving. Finally, the second method is implemented.Third, after combining model, parameter passing is a serious problem in model link running. The paper put forward a theory about dynamic link in model base and standard parameters. By adding a shield to the models stored in the model base, the author realizes standardization of models and an implementation scheme is given.Finally, combined model used to be an efficient way to resolve complex decision problems in IDSS. Whatever model combination technique is adopted, model chains generated in model combination process increases rapidly as the number of models in model bases increases so as to form a bottle-neck in problem solving speed. Considering the situations above, the author has put forth a distributed problem resolving model combination based on multi-agent technique, which adopts distribution technique to resolve problems in a parallel way to enhance the efficiency of problem-solving using combined model to ensure real-time decision process.