논문 해외 국제일반학술지 An algorithm for the stochastic vehicle routing problem using reinforcement learning
- 학술지 구분 국제일반학술지
- 게재년월 2002-01
- 저자명 임준묵공동(교신)
- 학술지명 Mathematical Decision Making under Uncertainty
- 발행처명 Research Institute for Mathematical Sciences
- 발행국가 해외
- 논문언어 외국어
- 전체저자수 3
홈페이지논문 초록 (Abstract)
In recent years, Artificial Intelligence techniques, like Deep Reinforcement Learning (DRL), have been used to propose solutions to complex computational problems. Vehicle Routing Problems are clear examples of that. Sometimes these kinds of problems consider uncertain information on their formulations, in what is known as Stochastic Vehicle Routing Problems (SVRPs), which are particularly useful to model real environments. In this work, the generic aspects and some research work that allowed advances in the state-of-the-art of SVRP’s solutions with DRL are discussed. The study conducted shows that Multi-Layer Perceptrons, for action value function modeling, are the most widely used DRL setting. However, it is necessary to venture into the use of other state-of-the-art approaches, such as models based on Multi-Head Attention trained with recent Actor-Critic algorithms.