A Study Of Task Sequencing And Target Allocating In Course Of Action

In modern war, COA is to find a good plan of actions, which is to get a reasonable combination of sub-actions and allocate the resources to the sub-actions rationally, in order to achieve maximum operational effectiveness. The methods of COA plays an important role in the field of situation assessment、production control, traffic control and the job-shop scheduling. An COA can often determine the performance of the plan, even the success or failure of the war. The paper mainly modeled and solved the problems of tasks sequencing and targets allocating of COA. The paper has done at least the following works:1. Task sequencing is divided into two cases, one is a number of different military troops through a checkpoint, we need to find a plan about the passing, the other is a plane bombing multiple targets, we need to find a plan about how to bomb the targets. In this paper, the problem is abstracted as a multi-objective optimization problem with constraint functions, the objective functions are determined by reliability, cost, time consumed of resources, and the constraint functions are determined by the precedence and resource availability. We define a uniform crossover, part crossover operator and two mutation operators, and a method based on evolution, a method based on multi-agent evolution, a method based on multi-agent co-evolution are proposed to solve the problem. The simulation demonstrated that these three methods can solve the problem good and receive a reasonable actions sequence.2. Task allocating in the way of one-to-one is mainly about the targets allocating. In this paper, the problem is abstracted as a Assignment Problem and solved by a method based on Hungarian algorithm and evolutionary algorithm. The result of allocation is represented by a matrix which element is "0" or "1", the "1" represents that the target which number is the number of the column is allocated to the plan which number is the number of the row. The matrix of the result is restrained that every column and every row has only one "1". The speed of Hungarian algorithm is high, but the algorithm can’t get a good performance on High-Dimensional problem. The evolutionary algorithm can resolve the High-Dimensional problem, but it costs more time. This paper combines these two methods to solve the problem, and the simulation demonstrated that this method is useful and reasonable.3. The problems always met in the allocating process include UAV joining the battlefield、UAV out of the battlefield、unstable of the result. This paper put forward solutions to these questions and experiments showed that these methods are effective. Also, this paper makes some research on the problem of task allocating in the way of more-to-more caused by the UAV joining the battlefield or out of the battlefield.This research is supported by the Basic Research Fund of Xidian University(No. JY10000902033).