Indeed in spite of the fact that methods of AI problems solving are developed for a long time, in our opinion only several methods can be used as theoretical base for AI problems solving:
- statistical theory of recognition;
- structural methods of recognition;
- logical approach to solving AI problems (fuzzy logic, expert systems);
All premises are appeared from other areas. Universal theory is not exists. But, for example, genetic algorithms and neuro nets allow effectively solving of optimization problems for AI researching.