Abstract:

A robust optimization mathematical model and an enhanced memetic algorithm (MA) for space launch mission scheduling are proposed to address the practical challenges in space launch mission scheduling under the influence of multiple uncertainties such as equipment failure, process rework due to quality issues, and technological changes. A robust optimization scheduling model for space launch missions is built by integrating the makespan under the worst-case scenario, the average maximum makespan, and the average deviation of process start times. A Q-learning enhanced memetic algorithm (QMA) is proposed, which can dynamically select parent solutions based on the population state. Simulation experiments demonstrate that QMA outperforms the traditional MA by an average of 1.02% in the comprehensive robustness metric, verifying its advantage in balancing scheduling efficiency and scheme robustness, and providing an effective decision-support tool for highly reliable space launch mission scheduling.

Key words: space launch mission scheduling, scheduling under uncertainty, robust optimization, memetic algorithm, Q-learning enhanced