Optimization and visualization in many-objective space trajectory design

Abstract

This work optimizes the thrusting profile of a low-thrust spacecraft propelled by an ion engine to raise from Earth’s low orbit to the vicinity of the Moon. The orbital raising phase is divided uniformly into sixteen sections, of which the first six are set to full propagation to escape early from the radiation belts, and the profiles of the other ten sections are subject to optimization together with the propagation start date and the spacecraft’s initial mass. Each section is defined by three variables. Thus, the optimization problem consists of thirty-two variables. Four objective functions are considered, namely the operation time of the ion engine system, time to reach the Moon, maximum eclipse time, and the initial mass of the spacecraft, subject to various constraints. We use the many-objective optimizer named Adaptive ε -Sampling and ε -Hood (A ε S εH) to search for non-dominated solutions, analyze the trade-offs between variables and objectives, and use a method called visualization with prosections to gain insights into the problem and to analyze the dynamics of the optimization algorithm.