Анализ Результатов Применения Алгоритма Муравьиной Колонии В Задаче Поиска Пути В Графе При Наличии Ограничений

The article describes an ant colony optimization algorithm applied to the solution of combinatorial optimization problems by the example of the problem of finding the shortest path in the graph. A number of modifications compared to the classical mathematical model M. Dorigo, which reduce the computational cost in finding solutions and improve its quality during parametric optimization (metaoptimization), are described. Proposed a method for comparing the quality of decisions based on the calculation of a number of indicators (average quality of decisions mark, deviation from optimal decision quality, probability of find decision, probability of find optimal decision) using in the process of comparison of different known heuristic methods. The results of comparing the quality of solutions with known methods for graphs with different numbers of nodes (different size of a problem) and different densities are given.