Automatic Time Allocation For Lijiang 2.4m Optical Telescope

With the development of astronomy,the demand for astronomical equipment is increasing rapidly.However,the equipment is sophisticated and expensive,thus the effective utilization of astronomical equipment is of great importance.So far,time allocation of optical telescopes is performed manually,which is time-consuming and inefficient.Moreover,time allocation of optical telescope is a NP-hard problem with multi-objective optimization.It is almost impossible for a human planer to find the optimum solution.So it is necessary for the automatic time allocation of optical astronomical telescope.Lijiang 2.4-meter telescope is one of the largest diameter general optical telescopes in East Asia,which has high resolution and precision in comprehensive performance.In this paper,we design a feasible and effective time allocation strategy for the 2.4-meter telescope to maximize its usage.In our work,we model the time allocation problem of 2.4-meter telescope and formalize it as a constraint optimization problem.At first,we improve the algorithm under Spike framework according to the time allocation problem.Then we propose a hybrid heuristic algorithm,which can be divided into two stages,to improve the time allocation.In the first stage,genetic algorithm is used to search the scheme with maximal scientific value.There are two features in this stage.Firstly,the weighted random strategy and greedy strategy applied to initialization can promote the diversity and quality of the initial population.Secondly,the two designed crossover operations are combined in our algorithm.It encourages the incorporation of useful gene segments.In the second stage,simulated annealing algorithm is utilized to increase the average continuity of users' observation as well as to maintain the scientific value.This can effectively avoid trapping in local optimum problem.The experiments are performed on two datasets: the real application data from 2.4-meter telescope and the analogue data sets.The experimental result verifies the feasibility and effectiveness of our algorithm.