Leader Reward for POMO-Based Neural Combinatorial Optimization