Multi-objective Bacterial Foraging Optimization Algorithm Based on Parallel Cell Entropy for Aluminum Electrolysis Production Process

Abstract

Environment-friendly aluminum electrolysis production process has long been a challenging industrial issue due to its built-in difficulty in optimizing numerous highly coupled and nonlinear parameters. This paper presents a multi-objective bacterial foraging optimization (MOBFO) algorithm to find optimal solutions that maximize the current efficiency and minimize the energy consumption and the production of perfluorocarbons (PFCs). Our method can be viewed as an enhanced version of the bacterial foraging optimization (BFO) in solving multi-objective optimization (MOO) problems (MOPs). We first propose a task-oriented optimization framework and model, and then parallel cell entropy and its difference are introduced to evaluate the evolutionary status of the Pareto solutions in a new objective space called parallel cell coordinate system (PCCS). In particular, the Pareto-archived evolution approach (PAEA) and the adaptive foraging strategy (AFS) are applied to balance the convergence and diversity of the Pareto front in the optimization procedure. Compared with traditional approaches, MOBFO not only increases speed of convergence toward the Pareto front, but also improves the diversity of the obtained solutions. Extensive experiment results on numerous benchmark problems and real-world aluminum electrolysis production process validated our proposed method’s effectiveness.