Design optimization under uncertainty is one of the most important topics in complex engineering systems. For instance, in engineering design optimization of gas engines (Fig.1), we may have to consider:
- Uncertainty due to uncontrollable variations
- Multiple disciplines
- Improved computational efficiency
Fig.1 Tolerance design of gas engines
Another example is system reliability. People rely on systems more and more for manufacturing and production, and without high reliability, systems will undergo frequent failures (Fig.2), which end up with huge economic loss and affect society as a whole.
Fig.2 System failures
In order to improve system reliability, our goal is to develop efficient robust optimization approaches and multi-disciplinary optimization methods with applications on various kinds of complex engineering systems.
Basically optimization is to find “best available” values of some objective function defined in a domain consisting of some constraints, and robust optimization is to find “best available” values when uncertainty exists (variations in variables/parameters) (Fig.3).
Fig.3 Multi-disciplinary optimization
We focus on single-disciplinary and multi-disciplinary optimization strategies incorporating robust optimization design into complex systems. The areas are (but not limited to):
- Single-objective Robust Optimiztion (RO) based on Sequential Quadratic Programming (SQP)
Hybrid multi-objective RO based on Differential Evolution (DE)
RO considering model uncertainty
Deterministic Multi-disciplinary Optimization (MDO)
System configuration improvement
Yanjun Zhang, Tingting Xia, JizhouZhang, Zixi Han