Reseach On The Method Of Military Planning Optimum Based On Hierarchical Task Network

Military planning is one of the most important steps in the whole process of military decision-making and commanding. Because of the complexity of the elements which tightly invovled with planning, such as situations, actions, resourses and so on, it appears extremely hard to find an effective approach to generate and seek an effective feasible plan. In the research domain of this problem, a planning menthod HTN has been introduced for a long time, which aimes to search the feasible blue print though the completely decomposing the missions and clearing up the conflicts. During the process of missions decomposition in the real world, due to the decomposition methods of mission nodes are not exclusive, the increasement of the decomposition methods of compound missions will definitely cause an exponential increasement of the number of the decomposition trees. If resourse scheduling is introduced into the proccess at this time, the same as considering the optimum assignment of resources during the mission decomposing, it will have to add algorithms for resource optimum in order to find a plan with optimum resource assignment from a mass of decomposition trees. It is doom to be a NP problem apparently. Aimed at this problem, this paper sheds a little glimmer on bringing forwards a structure framework, a process framework for optimum searching and an algorithm flow, all based on HTN planning methods and a optimum searching algorithm ties to Messy Genetic Algorithm, which solve the scheduling problem of resource optimum assignment during the mission decomposition.In the paper, based on introducing the summarization, formalizationand and existing algorithms, the structure framework integrates HTN planning and resource scheduling. The framework is comprosed of Mission Layer, Planning Layer, Scheduling layer, Regulation Layer, Resource Layer, Application Layer, and the complex connection relationships among them. Depend on the structure framework, the process framework describes the principium and process of the integration of planning and scheduling, which is carried out by four steps, describing the mission as task network based on the regulation, generating initial decomposition trees, introducing resource optimum algorithm, and figuring out the final generation course of actions. In the description of the algorithm, the system initialize the initial state, network task, mission, regulation and resource. The paper chooses GA as an example to show this process.Afterwards, the property of task network decomposing trees is systematically analyzed, which proves that different decomposing trees that belong to one task network exchange subtrees which both satisfy unique conditions, will definitely get the decomposing trees belong to the same task network. Based on this property, this paper brings forwards an HTN planning optimum searching algorithm based on messy chromosome genetic algorithm, to insure best resource assignment will be find rapidly. For this new algorithm, we design an emulational experiment to analyze and inprove the validity and applicability. Finally, a case of combined disembarkation is shown to check the algorithm brought forwards.