Roomenv V2

RoomEnv-v2 is a complex environment compatible with Gymnasium, designed to challenge reinforcement learning algorithms through a partially observable Markov decision process (POMDP). This environment leverages a knowledge graph-enhanced human memory system to improve the algorithm's decision-making capabilities in dynamic multi-room scenarios. Users can adjust parameters such as room layout, number of objects, and transition probabilities to create a variety of experimental settings, thereby evaluating and optimizing the agent's exploration and reasoning strategies.