An application of parallel computation to Collaborative Optimization

Multidisciplinary Design Optimization (MDO) has evolved as a discipline which provides a body of methods and techniques to assist engineers in solving large scale design problems. There are many frameworks for formulating MDO problems. These frameworks can be broadly classified as single-level or bi-level formulations. Collaborative Optimization (CO) is one of the popular bi-level formulations to solve an MDO problem. There are numerous design optimization problems which are highly CPU time intensive and require a long simulation time. With the advent of cheaper and faster available PC’s, distributed parallel computer clusters have become very popular. These clusters provide large computing power and can be used to solve problems faster and more efficiently. This research is an attempt to take advantage of the computational power of parallel computers in the field of design Optimization. The robust design optimization of an Internal Combustion Engine has been formulated using CO and implemented using parallel computers. Considerable savings in Wall Time has been achieved. A generic strategy for solving similar problems has also been devised. A benchmarking program has also been developed to assess theoretical speedup for any problem size. This program uses the Collaborative Optimization framework and simulates a design optimization on distributed memory clusters