An Interactive Framework for Offline Data-Driven Multiobjective Optimization

Abstract

We propose a framework for solving offline data-driven multiobjective optimization problems in an interactive manner. No new data becomes available when solving offline problems. We fit surrogate models to the data to enable optimization, which introduces uncertainty. The framework incorporates preference information from a decision maker in two aspects to direct the solution process. Firstly, the decision maker can guide the optimization by providing preferences for objectives. Secondly, the framework features a novel technique for the decision maker to also express preferences related to maximum acceptable uncertainty in the solutions as preferred ranges of uncertainty. In this way, the decision maker can understand what uncertainty in solutions means and utilize this information for better decision making. We aim at keeping the cognitive load on the decision maker low and propose an interactive visualization that enables the decision maker to make decisions based on uncertainty. The interactive framework utilizes decomposition-based multiobjective evolutionary algorithms and can be extended to handle different types of preferences for objectives. Finally, we demonstrate the framework by solving a practical optimization problem with ten objectives.