Non-Linear Constrained GRG Optimization under Heterogeneous Parallel-distributed Computing Environments

VAZQUEZ G.E.; RAINOLDI R.; BRIGNOLE N.B.

Computer-Aided Chemical Engineering

Elsevier

Año: 2000; p. 127 - 132

In this work, we have designed and implemented a parallel version of the GeneralisedReduced Gradient optimisation method (GRG), which was especially devised for efficientprocessing on a network of heterogeneous workstations (NOWs). The core parallel routines dealwith simultaneous constraint evaluation and the calculation of gradients for both the objectivefunction and the constraints. A hybrid model for task scheduling that minimises idle time andconsiders the heterogeneous nature of the processors was proposed. As to performancecomparisons, a modified speed-up metric that takes into account heterogeneity was employed.Significant time improvements were obtained for both academic and industrial examplescorresponding to process-plant units. The best results were attained for large-scale problems orwhen the functions to be evaluated were costly.