Decision making in the organizational and technical systems is complicated by the risk of various events emergence in the systems with a given probability. The article describes a method for supporting decision making under risk on the basis of integration of evolutionary, multi-agent, simulation models and numerical methods. In the course of the method's work the following steps are performing: a set of alternatives solutions is formed;  transition to a one-criterion alternatives evaluation is fulfilled by using linear convolution; transition to the system behavior certainty is fulfilled with the use of numerical optimization Bayes-Laplace criterion. The developed method is applied to solving the projects scheduling problem; the results agree with the problem solution that has been performed by using genetic algorithm method under certainty.

Keywords: decision making risks, multi-agent modeling, numerical methods of decisions optimization, scheduling problem, evolutionary modeling