|
Abstract
|
Currently, optimal decision-making problems in the real world have two types of complexity that limit the use of classical methods of single-objective planning in conditions of certainty: «several conflicting objectives instead of one objective function» and «several uncertainties in the parameters of almost any model». To solve the first problem, special methods of multi-objective decision-making were developed, and to solve the second problem, various methods of modeling and solving uncertain problems were developed, for example, such as stochastic (random) planning and robust (sustainable to the action of uncontrollable factors) planning. It should be noted that uncontrollable factors can arise both at the stage of production of products and at the stage of their exploitation. Due to the fact that in recent years there has been quite often the simultaneous use of a relatively large number of methods for achieving objectives in conditions of various uncertainties has taken place. At the same time, with the aim of successful search for optimal conditions of practical implementation, the basic principles of their division into two general categories of optimization methods have been developed: random multi-objective optimization and robust multi-objective optimization.
|