Optimal scheduling for location geosynchronous satellites refueling problem

• Published in: Acta astronautica 2019-10, Vol.163, p.264-271
• Main Authors: Zhang, Tian-Jiao, Yang, Yi-Kang, Wang, Bao-Hua, Li, Zhao, Shen, Hong-Xin, Li, Heng-Nian
• Format: Article
• Language: English
• Subjects: Algorithms; Ant colony optimization; Fuels; Geosynchronous satellites; Heuristic methods; Location-routing problem; Mission planning; Multiple fuel tankers; Multiple servicing vehicles; Numerical experiments; on-orbit refueling; Production scheduling; Refueling; Routing; Satellites; Schedules; Scheduling; Sequences; Tanker ships; Tankers; Vehicles
• ISSN: 0094-5765; 1879-2030
• DOI: 10.1016/j.actaastro.2019.01.024

This paper addresses the scheduling problem arising from refueling multiple geosynchronous (GEO) satellites with multiple servicing vehicles [many-to-many (M2M) refueling]. The problem is defined by a set of potential fuel tanker locations, a homogeneous fleet of servicing vehicles with limited capacities, and a set of fuel-deficient GEO satellites with known fuel demands. The objective is to open a subset of fuel tankers, to assign GEO satellites to these tankers and to design vehicle servicing sequences, in order to minimize the total mission costs. To achieve such an economical refueling strategy, the fuel tanker location and routing decision are required to be determined simultaneously. It is shown that this problem can be formulated as a location-routing problem (LRP) which is obviously NP-hard. In order to solve it, an ant colony optimization (ACO) metaheuristic that features a special solution representation scheme is proposed. The proposed algorithm is then tested on two types of instances depending on whether the cost for opening a fuel tanker is considered or not. Numerical experiments show that both the consideration of the fuel tanker opening costs and their locations do affect the refueling schedule. Furthermore, the proposed algorithm outperforms the existing approaches in the M2M refueling mission design, and it is able to detect the optimal refueling strategy on all instances containing 85 GEO targets with 15 potential fuel tankers while previous works stopped at 15 satellites and 2 fuel tankers. •Scheduling of the multiple satellites refueling mission is presented.•Servicing vehicle routing and the fuel tanker location are considered simultaneously.•An ant colony optimization based solving methodology is proposed.•A special solution presentation scheme is applied to improve the computational burden.