Optimal re-entry trajectory design by interactive multiobjective optimization with parallel programming

This paper addresses the problem of atmospheric re-entry trajectory optimization using a multiobjective formulation which permits complete trade-off between different performance criteria while reducing the number of constraints, in favor of both increased robustness and reduced computing time. The difficulty inherent to the definition of the multiobjective functional is solved by means of an interactive algorithm. Due to tile fact that solutions corresponding to different weight vectors are calculated in parallel, the decision maker can compare the results at any iteration. This allows mid-course corrections on the partial results as undesired solutions are discarded and the released processes are re-utilizcd to search in the vicinity of preferred points. The results approximate well those generated by a single-objective formulation where all the performance criteria but one are treated by means of inequality constraints. Up to four objectives are simultaneously considered in the optimization, namely crossrange, heating rate peak, total heat load and control cost.