Комбінаторика в задачах штучного інтелекту

Abstract

Для деяких задач штучного інтелекту з використанням теорії комбінаторної оптимізації побудовано математичні моделі. Показано, що в задачах цього класу комбінаторні конфігурації можуть бути як аргументом цільової функції, так і вхідними даними.

Для некоторых задач искусственного интеллекта с использованием теории комбинаторной оптимизации построены математические модели. Показано, что в задачах этого класса комбинаторные конфигурации могут быть как аргументом целевой функции, так и входными данными.

The problems of the artificial intelligence are difficult by nature and are not always amenable to formalization. A lot of the applied problems of this class are reduced to the problems of the combinatorial optimization. This is because their vast part requires sorting of variants. A combinatorial nature is the characteristic of the search problems. The design methods do not always explain the search nature of the artificial intelligence problems. The detailed analysis of the problems of this class shows that the argument for the objective function is the different types of the combinatorial configurations. The method of creating the artificial intelligence with the use of the combinatorial optimization theory is represented. An objective function and defined argument for the combinatorial configurations of different types are formulated. As the system analysis shows, in the problems of this class the combinatorial configurations can be the argument for the objective function and input data. The use of the combinatorial optimization theory allows to set the combinatorial nature, to formulate the objective function, to identify the characteristics signs, which establish the similarity of the artificial intelligence problems. The expounded researches allow the identification the uncertainty cause of different kinds, which arises up in the process of their decision, and explain the nature of the input data vagueness.