Sample path sharing in simulation-based policy improvement

Simulation-based policy improvement (SBPI) has been widely used to improve given base policies through simulation. The basic idea of SBPI is to estimate all the Q-factors for a given state using simulation, and then select the action that achieves the minimal cost. It is therefore of great importance to efficiently use the given budget in order to select the best action with high probability. Different from existing budget allocation algorithms that estimate Q-factors by independent simulation, we share the sample paths to improve the probability of correctly selecting the best action. Our method can be combined with equal allocation, Successive Rejects, and optimal computing budget allocation to enhance their probabilities of correct selection as well as to achieve better policies in SBPI. Such improvement depends on the overlap in reachable states under different actions. Numerical results show that with such overlap, combining our method with equal allocation, Successive Rejects and optimal computing budget allocation produces higher probability of selection as well as better policies in SBPI.

Authors

• Di Wu
• Qing-Shan Jia
• Chun-Hung Chen

Keywords

• Resource management
• Q-factor
• Aggregates
• Optimization
• Computational modeling
• Estimation
• Sample Paths
• High Probability
• Numerical Results
• Probability Of Selection
• Independent Simulations
• Budget Allocation
• Equal Allocation
• Reachable States
• State Space
• Transition Probabilities
• Optimal Policy
• Markov Decision Process
• Candidate Solutions
• Simulation Optimization
• Bandit Problem
• Discrete event dynamic system
• simulation-based optimization
• optimal computing budget allocation