Search

The purpose of this paper is to seek the suitable local objectives for each satellite to optimal the global mission allocation strategy with a global utility function. This paper uses a game-theoretical formulation for a multi-satellite system in which the satellites are viewed as a unique unit with their self-interests. To solve the problem, the first is to identify the utility functions for individual satellite, align them to form the global utility function which can represents the allocation requirements. The second is to design the suitable negotiation mechanisms that can be equipped on each satellite so they can pursue the optimization by their own interest. The Utility-based Regret play, the Smoke Signal play, and the Broadcast-based play are proposed as negotiation mechanisms for the team to cooperate under the distributed and decentralized system structure. The simulation results illustrate that using these mechanisms can help this multi-satellite system reaches a near-optimal allocation profile. The effectiveness of proposed mechanisms are demonstrated by comparing their simulation results with several existing mechanisms under different scale of participant number. [ABSTRACT FROM AUTHOR]

Space missions have traditionally been controlled by operators from a mission control center. Given the increasing number of satellites for some space missions, generating a command list for multiple satellites can be time-consuming and inefficient. Developing multi-satellite, onboard mission scheduling & planning techniques is, therefore, a key research field for future space mission operations. In this paper, an improved Genetic Algorithm (GA) using a new mutation strategy is proposed as a mission scheduling algorithm. This new mutation strategy, called Hybrid Dynamic Mutation (HDM), combines the advantages of both dynamic mutation strategy and adaptive mutation strategy, overcoming weaknesses such as early convergence and long computing time, which helps standard GA to be more efficient and accurate in dealing with complex missions. HDM-GA shows excellent performance in solving both unconstrained and constrained test functions. The experiments of using HDM-GA to simulate a multi-satellite, mission scheduling problem demonstrates that both the computation time and success rate mission requirements can be met. The results of a comparative test between HDM-GA and three other mutation strategies also show that HDM has outstanding performance in terms of speed and reliability. [ABSTRACT FROM AUTHOR]