Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications

This work proposes a novel framework that combines logic programming with large language models (LLMs) to automatically generate Markov Decision Processes (MDPs) for complex robotic tasks. By integrating symbolic reasoning with the generative capabilities of LLMs, the framework enables the creation of structured MDPs that can effectively model the uncertainties and dynamics of real-world robotic environments. The proposed approach is validated through extensive simulations and real-world experiments, demonstrating its effectiveness in enhancing robot decision-making and task execution.

Recommended citation: E. Saccon, D. De Martini, M. Saveriano, E. Lamon, L. Palopoli and M. Roveri, "Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications," in IEEE Robotics and Automation Letters, vol. 11, no. 2, pp. 1770-1777, Feb. 2026
Download Paper