Distribution affects company cost and customer satisfaction. Distribution network design is a strategic concern for nearly all companies, and must be periodically reviewed to ensure fitness to today's rapidly changing business environment. Unlike most of previous research, this paper investigates a highly complicated and realistic distribution network, and develops a multi-objective model to achieve an effective redesign solution. Reconfiguring cost, service differentiation, multiple products, and optional capacity levels are simultaneously considered. A determinant code-based encoding structure is designed to represent the complicated multi-objective optimization problem, and a genetic algorithm (GA)-based procedure is developed as a solution. The paper provides an application example using actual data. Finally, a comparison between the empirical results and those of the Prüfer-based GA demonstrate the efficiency of the proposed GA.