The Reply Plan Of Adversarial Plan

Planning is a special case of automatic solution problem. It's also an important sub-field in AI. The main research in AI is use artificial method and technology to imitate, extend and expend human intelligence and ability. This can make machine have intelligence and ability as human. What we most concern is how the intelligence system deduces a goal properly, after that, how it makes an action sequence which attempt to achieve the goal. This process is called planning.Intelligence plan and plan recognition are a subject which involve knowledge express, knowledge inference, nonmonotonicity logics, cognition science and so on. The research on intelligence plan and plan recognition can make profound meanings in AI.Enlightened by biology stress reaction, neurobiology of learning and memory, cell and molecular biology, the research on reply plan of adversarial in this thesis is in several aspects, as follow:1. The reply plan of adversarial plan model: define a conception of stress reaction, set up a stress reaction system model, study the stress calculate algorithm. The stress reaction system model and its relative algorithm can make the adversarial and reply plan react as rapidly as biology stress reaction.2. The start mechanism on reply plan: define a conception of reply plan sign, it can make plan recognition (plan recognition system) recognize an integer plan of opponent and attack goal from the partly and dispersed actions; and can forecast potential danger from opponent's succeed actions.3. The solution search on reply plan: use the reality word in dynamic system, this thesis builds a rough dynamic system to do the solution search under all the possible state.4. The establish of reply plan stock: simulate the main learn-memory nerve circuit of human, use neuron group as a independent union set up a distribution neural network; use neuron group stimulus value as the signal of learn and memory. This can make the reply plan have the permanent resistant capacity to an adversarial plan which has been recognized.5. Use catastrophe theory defines a function—stress reaction catastrophe function(SRC)—to reply a class of adversarial plan which makes up by many non-adversarial actions.Use AI and intelligence plan technique to imitate biology stress reaction and learn-memory nerve circuit can make intelligent has the characters: resist on real-time, start on time, reply effectively and defend permanently. This research has an extensive applied prospect in contest robot, game role design and multi-agent system and so on.