Author(s)

N. Bouhmala, K. Ghedira & H.H. Naegeli

Abstract

Large meshes computations arise in many large-scale scientific and engineering problems, including finite volume methods for computational fluid dynamics, and finite element methods for structure analysis. If these meshes have to be solved efficiently on distributed memory parallel processors, a partitioning strategy should be designed so that on the one hand, processors have approximately equal work to do, and on the other hand inter-processor communication is minimized. In this paper we introduce a distributed optimization model combining Multi-agent systems and simulated annealing for the mesh partitioning problem. 1 Introduction Many large-scale computational problems are based on unstructured computations domains. Among such problems, unstruct

Keywords