Open AccessThe modern methods to solve the practical tasks related to designing the information security systems in automated systems of different purpose rather often use models of the theory of games. The work, for resource distribution between securable objects, formulates a task of the two players’ game (a defender and a forward) with the zero-sum. The possible loss of a defender is used as a score. The statement of problem for each player, when holding the other player’s solution, is the task of linear programming with one restriction. For the solution of similar tasks the paper offers a linear programming method.
For a defender to search a solution, it is offered to use a saddle-point criterion. If the saddle point is not found, a criterion of the guaranteed result is used. For search of the saddle point, two algorithms are offered. Originally, an optimization algorithm, based on the sequential solution of the optimization tasks by each player, seeks the saddle point. If the first algorithm does not find the saddle point, there is a developed search algorithm of the saddle point based on the support of equal security of some objects of a securable system. For the support of equal security of some objects an offer is to solve a system of linear equations. For algorithm are formulated and proved sufficient conditions that the received solution is the saddle point.
If both algorithms did not find the saddle point, for a defender to obtain the solution, using a principle of guaranteed result is offered. The search algorithm of solution has been developed to provide the guaranteed result based on the algorithm of support of equal security in which it is possible to use the initial quality score or, in case there is no information about forward ‘s resources, the heuristic scores.
Results of the task solution of resource distribution with the given basic data, as well as test results of the developed algorithms are provided.