Approaches That Use Domain-Specific Expertise: Behavioral-Cloning-Based Advantage Actor-Critic in Basketball Games

Research on the application of artificial intelligence (AI) in games has recently gained momentum. Most commercial still use AI based a finite state machine (FSM) due to complexity and cost considerations. However, FSM-based decreases user satisfaction given that it performs same patterns consecutive actions situations. This necessitates new approach applies domain-specific expertise existing reinforcement learning algorithms. We propose behavioral-cloning-based advantage actor-critic (A2C) improves performance by applying behavioral cloning algorithm an A2C basketball games. The normalization, reward function, episode classification approaches are used with A2C. results comparative experiments traditional algorithms validated proposed method. Our method using solved difficulty