The Knowledge-based Genetic Algorithm to the Disasters Monitoring Task Allocation Problem

As increasing frequency of geological disasters occurred, the size of monitoring tasks and imaging satellites is rising fast, which increase the computational complexity of traditional task scheduling algorithms vastly. Meanwhile, satellites from different departments could not be put into an algorithm framework to optimize. In fact, noticing the unfitness of traditional multi-satellites scheduling algorithms in disasters monitoring, pre-allocation of tasks to satellites becomes an effective means. In this paper, we proposed a novel knowledge-based genetic algorithm(KGA) based on analysis of specialties and operational constraints of tasks' allocation. The KGA is constructed in a three-dimensional architecture based on genetic algorithm and problem's character and users' priority. Two extended heuristic approaches are applied to produce initial individuals, and the Components Combination Knowledge is employed to decide a suitable broken position for operations of crossover and mutation, and the Partial Replacement Procedure is implemented to maintain population diversity. The simulation and experimental results show the feasibility and adaptability of KGA for disasters monitoring tasks allocation problem, compared with traditional task allocation algorithms.