A Knowledge Recognition Based Planning Method

Intelligent planning means that an agent seeks a sequence of actions for a specified goal state that can solve a particular problem, starting from the initial state of the problem. Intelligent planning is also known as automatic planning, and it is an important research area in artificial intelligence. It is also considered as a cross multi-field discipline, mainly including knowledge representation and reasoning, human-computer interaction, cognitive science and many other fields. This paper proposes two methods to solve the corresponding practical application problems based on intelligent planning technology.In order to solve the planning problem of low efficiency and high cost in intelligent systems, this paper proposes a BDI(Belief, Desire, Intention) planning method based on knowledge recognition, to achieve the reuse of planning knowledge and improve the planning efficiency. By using plan recognition model proposed based on the agent programming framework and following the behavior model represented by the workflow-net, the patterns of the sequence of behavior are discovered and derived. Furthermore, the planning structure proposed including planning body and the context, recognized by BDI Agents, can be transferred from the workflow-net. In this paper, a new method of BDI planning based on knowledge recognition is proposed, and this method is a new initial attempt in the intelligent planning methods. The experimental results prove the feasibility of the knowledge recognition based BDI planning method. and they have some practical significance.Goal recognition is viewed as a special plan recognition, and it is the supplement and improvement for the planning knowledge. The goal recognition method based on Agent programming framework is proposed. On the basis of the planning recognition, the BDI goals can be recognized following that whether the goal library exists or not. Through the experiment, the important factors on the performance of the algorithms are analyzed, namely the expansion node in the workflow-net, choice branching factor and parallel branching factor. The experimental results are concluded. The BDI planning based on the knowledge recognition is a plan derived for BDI agent oriented programming paradigm, and the process for the recognition of BDI plan is also the process of learning BDI agent program.In order to solve the problem of cooperative convergence of selfish nodes in the network system, this paper proposes a cooperative strategy based on intelligent planning, which combines game theory and planning knowledge to analyze the behavior strategy, and further make decisions to achieve the goal state of the cooperation convergence among nodes. In this paper, the convergence of the interactive cooperation strategy of the network nodes is analyzed under different static topologies. The experimental results show that under the certain conditions, the performance of the proposed method is better than other deterministic update strategy planning methods, including IBN, IBS and WSLS. The research results of this paper provide a new way and solution of the practical problem to promote the research of intelligent planning and plan recognition.To draw an conclusion, intelligent planning is a problem solving technique, BDI planning method research based on knowledge recognition(including the plan recognition and target recognition), is able to achieve the reuse of BDI programming knowledge and improve the efficiency of research and development of intelligent system; The cooperation strategy based on intelligent planning is able to direct and adjust the behavioral strategies of wireless sensor network nodes, achieving the goal state of the cooperation convergence among the selfish nodes. It also contributes to the performance and efficiency improvement of the network.