Methods of rational decision making in multi-agent systems for evaluating the effectiveness of innovations

Abstract

The work is devoted to research and development of new meta-heuristic methods and algorithms of rational choice, planning and optimization of management decisions. The search for a solution is carried out using a multi-agent approach, which can be used to evaluate innovative activities. The scientific novelty of the proposed multi-agent technology lies in the fact that intelligent agents act on the basis of a function of fuzzy preferences. An original method is proposed for determining fuzzy judgments of agents, which combines the approaches used today with fuzzy automatic control based on expert judgments and the idea of dividing the criteria space into areas proposed by the authors in the method of combined decision support. The scientific significance of solving such problems is manifested in the case of a high dimensionality of the vector criterion and optimized parameters of the domain model, as well as in the case of complex algorithmic rules of the business model of innovative development. For this purpose, a software prototype of an environment for integrating distributed agents in a single information space has been developed to search for management solutions that are close to optimal. The proposed approach, based on fuzzy judgments of agents, will optimize the development programs of innovatively active enterprises and solve scientifically significant large-scale tasks of scheduling.The work is devoted to research and development of new meta-heuristic methods and algorithms of rational choice, planning and optimization of management decisions. The search for a solution is carried out using a multi-agent approach, which can be used to evaluate innovative activities. The scientific novelty of the proposed multi-agent technology lies in the fact that intelligent agents act on the basis of a function of fuzzy preferences. An original method is proposed for determining fuzzy judgments of agents, which combines the approaches used today with fuzzy automatic control based on expert judgments and the idea of dividing the criteria space into areas proposed by the authors in the method of combined decision support. The scientific significance of solving such problems is manifested in the case of a high dimensionality of the vector criterion and optimized parameters of the domain model, as well as in the case of complex algorithmic rules of the business model of innovative development. For this...