Software

Adaptive Response Surface Method (ARSM)

 

ARSM is a global optimization scheme developed for design optimizations involving computation-intensive processes such as finite element analysis (FEA), computational fluid dynamics (CFD), simulation, and so on.  For detailed information, one can download the papers (Wang et al. 2001) and (Wang 2003).  ARSM is fully developed in the PDOL and its first version is completed in July 2001. It has been successfully applied to helicopter component design, industry silencer design, etc.  ARSM then leads to the development of  MPS as below.

 

Mode Pursuing Sampling (MPS) Method

 

As an improvement from ARSM, MPS is an efficient global optimization method for both constrained and unconstrained optimization problems (Wang et al. 2004).  Its discrete-variable version was also developed (Sharif et al. 2008).  The performance of MPS was studied in reference to genetic algorithms (Duan et al. 2008). MPS is considered a step-up from ARSM. Download Codes (.zip)

 

Pareto Set Pursuing (PSP) Method

 

PSP is an efficient multi-objective optimization (MOO) method.  It deals with problems with continuous, discrete, and mixed-variables (Shan and Wang 2005). Its performance is also compared with other widely-used MOO methods such as multi-objective genetic algorithms (MOGA), NSGA-II, FastPGA, SPEA2, etc. It is found that for a limited number of function evaluations, PSP generates much better Pareto Frontier sets, as compared with others (Khokhar et al. 2010). Download Codes (.zip)

 

OPTIP

 

In Summer 2010, MPS and its variation CiMPS, as well as PSP, are integrated into a new optimization tool, named Optimization Toolkit for Computationally Intensive Problems (OPTIP).  This tool has Graphic User Interface with many plots for the optimization process and outputs. In 2011, a discrete version of MPS, DMPS, is added. Download Codes (.zip)

 

Others:

 

Reliability-based Design Optimization

Collaboration Pursuing Method (CPM) for Multidisciplinary Design Optimization

Product Design and Optimization Lab  (PDOL)