Pareto-optimization of complex system architecture for structural complexity and modularity

Due to ever-increasing complexity of cutting-edge engineering systems, the need for managing structural complexity and modularity of such systems is becoming important. The complexity of the overall system architecture is mostly decided during the initial concept generation stage, when configurations of major modules within the system are determined. In this paper, we present a multi-objective optimization framework for (1) minimizing the variation in complexity allocation to individual modules, while (2) maximizing for the degree of modularity. The optimization framework was applied to a case study, where a trailing bogie system for railroad train was optimized for structural complexity allocation among individual modules and overall system modularity. The modularity maximizing decomposition is shown to induce a large variation in module-level complexity distribution with a small fraction of modules sharing a disproportionately large chunk of overall system complexity, while equitable distribution of module-level complexity leads to erosion in the degree of modularity achieved for the resulting system decomposition.