Pattern-based integration of system optimization in mechatronic system design.

Wenqiang Yuan,Yusheng Liu,Jianjun Zhao,Hongwei Wang
DOI: https://doi.org/10.1016/j.advengsoft.2016.03.005
IF: 4.255
2016-01-01
Advances in Engineering Software
Abstract:A new profile named is proposed to implement the selection of optimization method based on SysML.A pattern with mechatronic semantics is formalized for supporting the retrieval of feasible optimization methods.A general automated selection process of optimization method is proposed based on the optimization profile.The detailed retrieval process of optimization methods and the semantic similarity calculation are proposed. Model-Based Systems Engineering (MBSE) is a promising methodology for the design of complex mechatronic systems. There are some tools developed to support MBSE in complex system design and modeling. However, none of them has got the functionality of supporting system optimization. This work is precisely motivated by this gap and aims to develop effective methods to support automatic system optimization for MBSE. Specifically, a pattern-based method is proposed to support the integration of system optimization into mechatronic system design. In such a method, optimization problems, their solving methods and computation results are formally defined in each pattern based on the System Modeling Language (SysML). In addition, a model description scheme termed an optimization profile is proposed based on SysML to include the components for formalizing different kinds of optimization problems and optimization methods. After an optimization profile is created for an optimization problem, system optimization methods can be chosen automatically from the pattern library based on a semantic similarity evaluation. Then, optimization results are provided to users to support decision-making and the pattern library is updated using the relevant information obtained in this process. A system design example is used to demonstrate the feasibility and effectiveness of proposed methods.
What problem does this paper attempt to address?