Research On Multi-target Task Planning In JSHOP2

Java Simple Hierarchical Ordered Planner(JSHOP2)is widely used in intelligent planning problems because of its simple and efficient task decomposition network.Current JSHOP2 planner can complete task planning in an ordered mode or unordered mode,but ignores the priority of different targets,so it cannot handle multi-target task planning problems with priorities well.At the same time,JSHOP2 multi-target task planning usually produces multiple feasible solutions,so how to get the optimal plan needs further study.Aiming at the problems of multi-target task planning in JSHOP2,the thesis proposes the multi-target priority task planning method in JSHOP2 and the optimal planning scheme selection method based on multi-target operation cost in JSHOP2,so as to support the multi-target priority and obtain the optimal operating cost planning scheme.The main research results of the paper are as follows:1.Research on task planning based on multi-target priority in JSHOP2This research first builds a domain editor that supports multi-target priority to provide support for rapid and efficient multi-target priority construction.Then,based on the PDDL language,the multi-target priority is extended to provide the priority setting syntax.Finally,a task planning algorithm for static target priority and a task planning algorithm for dynamic target priority are proposed.Under the premise of ensuring task planning constraints,the target with higher priority can be completed first.2.Research on selection method of optimal planning scheme based on multi-target operation cost in JSHOP2This research first sets the operating cost for the atomic task based on the task characteristics,and then proposes an algorithm selection algorithm based on the operating cost.This algorithm calculates the total operational cost value of the planning program during the task iterative planning process to select.Then proposes a planning optimization algorithm based on operating cost,prunes the disadvantaged scheme based on the current planning state,reduces the execution time of multi-target task planning,and finally selects the optimal planning scheme for task planning problem.This paper constructs a simulation experiment based on logistics scenarios.The experiments show that the above algorithms can achieve reasonable results and high execution efficiency on the JSHOP2 planner,and can solve the problem of priority processing and scheme selection of multi-task planning in JSHOP2.