Processos de decisões de Markov

Abstract

This study explores Sequential Markov Decision Models, or Markov Decision Processes (MDPs), which are used for sequential decision-making in stochastic environments. With practical applications in areas like marketing and inventory management, MDPs help optimize adaptive decisions, such as advertising campaigns and inventory restocking, reducing costs and enhancing customer experience. The research is structured into chapters that provide a gradual approach to the topic. The first chapter introduces a probability foundation, essential for understanding Markov Processes. The second chapter focuses on Markov Chains, fundamental to comprehending MDPs. Subsequent chapters delve into non-homogeneous aspects of Markov Processes, detailing temporal variations and variable rewards, enabling the modeling of diverse phenomena. The work includes theoretical and practical examples to illustrate concepts and explores topics such as absorbing states and stopping problems in MDPs, making theory application more accessible and enhancing reader comprehension.