TycoonLE: A Jax reinforcement learning environment for long-horizon planning

TycoonLE is a JAX-based reinforcement learning environment for long-horizon economic planning where agents manage logistics operations including capital allocation, transport routes, cargo movement, and debt management. The environment features a fixed-shape interface compatible with JAX transformations, a replay UI for policy inspection, and recent updates including OpenTTD validation and agent evaluation tools. TycoonLE is accompanied by TycoonBench, a benchmark suite for comparing agent and model performance on planning tasks.