Alessandro Trapasso

Alessandro Trapasso

I am a Postdoctoral Researcher at the PSO Unit, Fondazione Bruno Kessler (FBK). I earned my PhD in Artificial Intelligence in September 2025.

About

My name is Alessandro Trapasso. I am a Postdoctoral Researcher at the Planning, Scheduling and Optimization Unit (PSO) at Fondazione Bruno Kessler (FBK), where I work on reinforcement learning for automated planning and scheduling within STEP-RL, an ERC Starting Grant led by Andrea Micheli. I earned my PhD in Artificial Intelligence from Sapienza University of Rome in September 2025.
I worked with the research groups RoCoCo Lab and WhiteMech, and my supervisors were Prof. Luca Iocchi and Prof. Fabio Patrizi.

During my PhD, I was a visiting researcher at the Universitat Pompeu Fabra, Barcelona, where I worked with Prof. Anders Jonsson.

I earned my MSc and BSc in Computer Science and Engineering at Sapienza University of Rome.
Since 2022, I have been a licensed Professional Engineer in Italy.

My research focuses on the integration of automated planning and reinforcement learning, with particular interest in model-based and multi-agent RL, non-Markovian rewards, reward machines, temporal logics and automata-based methods, concurrency, and coordination.

I also develop and maintain Multi-Agent RLRM, an open-source Python framework for multi-agent reinforcement learning with Reward Machines, including tools for specification, evaluation, and reproducible experimentation.
In addition, I led the development of the multi-agent planning module in Unified Planning.

I am passionate about Machine Learning, football, and chess.
I believe curiosity and continuous self-improvement make a difference in any field where you want to excel.
I am a positive person and I like to invest my time in what I love.

Selected Publications

IJCAI-ECAI 2026

Model-Based Reinforcement Learning in Discrete-Action Non-Markovian Reward Decision Processes

Alessandro Trapasso, Luca Iocchi, and Fabio Patrizi

arXiv
ECAI 2025

Concurrent Multiagent Reinforcement Learning with Reward Machines

Alessandro Trapasso and Anders Jonsson

Publisher
SoftwareX 2025

Unified Planning: Modeling, Manipulating and Solving AI Planning Problems in Python

Andrea Micheli, Arthur Bit-Monnot, Gabriele Röger, Enrico Scala, Alessandro Valentini, and others

Publisher

View all publications

Contact