A distributional view on multi-objective policy optimization
Abbas Abdolmaleki,u00a0Sandy Huang,u00a0Leonard Hasenclever,u00a0Michael Neunert,u00a0Francis Song,u00a0Martina Zambelli,u00a0Murilo Martins,u00a0Nicolas Heess,u00a0Raia Hadsell,u00a0Martin Riedmiller
Many real-world problems require trading off multiple competing objectives. However, these objectives are often in different units and/or scales, which can make it challenging for practitioners to express numerical preferences over objectives in their native units. In this paper we propose a novel algorithm for multi-objective reinforcement learning that enables setting desired preferences for objectives in a scale-invariant way. We propose to learn an action distribution for each objective, and we use supervised learning to fit a parametric policy to a combination of these distributions. We demonstrate the effectiveness of our approach on challenging high-dimensional real and simulated robotics tasks, and show that setting different preferences in our framework allows us to trace out the space of nondominated solutions.
