Multi-objective design optimization of a new space radiator

In this work different multi-objective techniques are used to the conceptual design of a new kind of space radiator. Called VESPAR (Variable Emittance Space Radiator), the radiator has an effective variable emittance which makes it able to reduce or avoid the demand for heater power to warm up equipment during cold case operations in orbit. The multi-objective approach was aimed on obtaining a radiator that minimize its mass while at the same time minimize the need for heater power during cold case. Four multi-objective algorithms were used: Elitist Non-dominated Sorting Genetic Algorithm (NSGA-II), Multi-Objective Genetic Algorithm (MOGA), Multi-Objective Simulating Annealing (MOSA) and Multi-Objective Generalized Extremal Optimization (M-GEO). The first three algorithms were used under the modeFrontier^® optimization software package, while M-GEO is a recently proposed multi-objective implementation of the Generalized Extremal Optimization (GEO) algorithm. The Pareto frontier showing the trade-off solutions between radiator mass and heater power consumption is obtained by the four algorithms and the results compared. An assessment of the performance of M-GEO on this problem, compared to the other well-known multi-objective algorithms is also made.