In this article a genetic algorithm (GA) based application for computer aided design and optimization of cam shapes used on internal-combustion engine camshafts is discussed. A shape optimization of cam cross-section is a multiobjective optimization problem of two-dimensional geometric shape in a heavily constrained environment.

A cam shape is parametrized using B-splines. By representing the cam shape in parametric form, the shape optimization problem is converted to a parameter optimization problem. The real valued control points of the B-splines are used as chromosomes of a real coded genetic algorithm. Each individual of the population consists of a set of 40 floating-point values which unambiguously describes a cam shape. Thus, the individuals of the population are actually the alternative cam shape designs in a parametric form.

The fitness-function of the genetic algorithm is based on evaluation of a simulated cam mechanism model. The alternative cam shape designs are first tested on simulated cam mechanism. Then the fitness-value for each individual cam is calculated on basis of the simulation results.

This article focuses on possibilities to implement parallelly or/and distributedly this kind of shape optimization system in order to speed up the time consuming optimization process. A distributed computation method in a local area network (LAN) is described. The method is based on a population maintained by a master process. Evaluation of fitness-values is distributed via a local area network to slave processes which work partially asyncronously with respect to the main process. The number of slave processes can be freely selected and also freely altered during the optimization process. Shared disk files are used as a logical interface between master and slave processes.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £56 +P&P)