An mTSP model for multiple telescope short-term scheduling

Abstract

Pulsar observation scheduling is a highly dynamic problem with uncertainty caused by scintillation and environmental interference. In the scheduling systems used by observatories today, this uncertainty is usually accounted for by using a short-term scheduling process. In order to observe as many pulsars as possible within a certain period of time, it becomes imperative for multiple telescopes or sub-arrays to observe multiple targets at the same time. However, there is little research on the speed improvement and cost-efficiency improvement of such “parallel" observation with multiple telescopes. In this work, we have designed a multiple telescope short-term scheduling simulation system, based on extending previously proposed scheduling heuristics using a single telescope pulsar scheduling simulator. Additionally, we propose a more complex multiple telescope short-term scheduling policy based on the multiple travelling salesperson problem (mTSP). Realistic datasets within the simulation environment are used to evaluate the proposed algorithms with multiple telescopes. The results show that multiple telescopes can significantly reduce the time to observe a set of pulsars. The previous single telescope heuristics adapt surprisingly well to the multiple telescope scenario and the novel mTSP policy is competitive, with potential for future improvements.