Continuous-time zero-sum games for markov decision processes with risk-sensitive finite-horizon cost criterion on a general state space

• Autores: Subrata Golui, Chandan Pal
• Localización: 3c Empresa: investigación y pensamiento crítico, ISSN-e 2254-3376, Vol. 11, Nº. 2, 2022, págs. 76-92
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Dialnet Métricas: 1 Cita
• Resumen

  • In this manuscript, we study continuous-time risk-sensitive finite-horizon time-homogeneous zero-sum dynamic games for controlled Markov decision processes (MDP) on a Borel space. Here, the transition and payoff functions are extended real-valued functions. We prove the existence of the game’s value and the uniqueness of the solution of Shapley equation under some reasonable assumptions. Moreover, all possible saddle-point equilibria are completely characterized in the class of all admissible feedback multi-strategies. We also provide an example to support our assumptions.