A system sensitivity analysis-based parallel method for optimization of product platform

In order to find the optimal product platform in a product family, a system sensitivity analysis-based parallel method for optimization of product platform (SSAPM) is proposed. The key idea of this method is to decompose the product platform optimization problem with a system-level optimization problem and several independent subsystem-level optimization problems by introducing coordination variables. The system-level optimization obtains the sensitivity information on the objective function about the coordination variables by calling the subsystem-level optimization, and adjusts the coordination variables according to the sensitivity information. Under product constraints and performance constraints, the subsystem optimizes each product so that the design variables of each product are as close as possible to the coordination variables. With the continuous iteration between the system and the subsystems, the coordination variables that can maximize the generality of the product family are finally found. The numerical test results show that SSAPM has strong optimization ability.

Automated summary

SSAPM is a parallel optimization method based on system sensitivity analysis, which decomposes the problem into system-level and subsystem-level optimization problems, and introduces coordination variables to find the optimal product platform. Through continuous iteration, the coordination variables that can maximize the generality of the product family are eventually found.