Dual-telescope observation plan execution: DRL-GR results outperform the greedy and random scheduling. Sequential processes are marked by red arrows. Parenthetical values denote the total theoretical reward scores. DRL-GR-1 plans (bottom left) achieve 33.9% higher task completion and 24.1% reward gain despite wider spatial distribution vs. the greedy benchmark.