MORL/D: Multi-Objective Reinforcement Learning based on Decomposition

in International Conference in Optimization and Learning (OLA2022) (2022)

Metaheuristics-based Exploration Strategies for Multi-Objective Reinforcement Learning

in Proceedings of the 14th International Conference on Agents and Artificial Intelligence (2022)

The fields of Reinforcement Learning (RL) and Optimization aim at finding an optimal solution to a problem, characterized by an objective function. The exploration-exploitation dilemma (EED) is a well ... [more ▼]

The fields of Reinforcement Learning (RL) and Optimization aim at finding an optimal solution to a problem, characterized by an objective function. The exploration-exploitation dilemma (EED) is a well known subject in those fields. Indeed, a consequent amount of literature has already been proposed on the subject and shown it is a non-negligible topic to consider to achieve good performances. Yet, many problems in real life involve the optimization of multiple objectives. Multi-Policy Multi-Objective Reinforcement Learning (MPMORL) offers a way to learn various optimised behaviours for the agent in such problems. This work introduces a modular framework for the learning phase of such algorithms, allowing to ease the study of the EED in Inner- Loop MPMORL algorithms. We present three new exploration strategies inspired from the metaheuristics domain. To assess the performance of our methods on various environments, we use a classical benchmark - the Deep Sea Treasure (DST) - as well as propose a harder version of it. Our experiments show all of the proposed strategies outperform the current state-of-the-art ε-greedy based methods on the studied benchmarks. [less ▲]

Automating the Design of Efficient Distributed Behaviours for a Swarm of UAVs

in IEEE Symposium Series on Computational Intelligence, Canberra 1-4 December 2020 (2020, December)

The usage of Unmanned Aerial Vehicles (UAVs) is gradually gaining momentum for commercial applications. The vast majority considers a single UAV, which comes with several constraints such as its range of ... [more ▼]

The usage of Unmanned Aerial Vehicles (UAVs) is gradually gaining momentum for commercial applications. The vast majority considers a single UAV, which comes with several constraints such as its range of operations or the number of sensors it can carry. Using multiple autonomous UAVs simultaneously as a swarm makes it possible to overcome these limitations. However, manually designing complex emerging behaviours like swarming is a difficult and tedious task especially for such distributed systems which performance is hardly predictable. This article therefore proposes to automate the design of UAV swarming behaviours by defining a multi-objective optimisation problem, so called Coverage of a Connected-UAV Swarm (CCUS), and designing a Q-Learning based Hyper-Heuristic (QLHH) for generating distributed CCUS heuristics. Experimental results demonstrate the capacity of QLHH to generate efficient heuristics for any instance from a given class. [less ▲]

Automated design of efficient swarming behaviours: a Q-learning hyper-heuristic approach

in GECCO '20: Genetic and Evolutionary Computation Conference, Companion Volume, Cancún, Mexico, July 8-12, 2020 (2020)