A Learning Algorithm for Risk-Sensitive Cost

• 成果类型: Article
• 署名作者: Basu, Arnab; Bhattacharyya, Tirthankar; Borkar, Vivek S.
• 署名单位: Indian Institute of Management (IIM System); Indian Institute of Management Bangalore; Indian Institute of Science (IISC) - Bangalore; Tata Institute of Fundamental Research (TIFR)
• 刊物名称: MATHEMATICS OF OPERATIONS RESEARCH
• ISSN/ISSBN: 0364-765X
• DOI: 10.1287/moor.1080.0324
• 发表日期: 2008
• 页码: 880-898
• 关键词: large deviations; SPECTRAL THEORY; DYNAMICS

摘要

A linear function approximation-based reinforcement learning algorithm is proposed for Markov decision processes with infinite horizon risk-sensitive cost. Its convergence is proved using the o.d.e. method for stochastic approximation. The scheme is also extended to continuous state space processes.