A parallel computing application of the genetic algorithm for lubrication optimization

Abstract

This study investigated the performance of parallel optimization by means of a genetic algorithm (GA) for lubrication analysis. An air-bearing design was used as the illustrated example and the parallel computation was conducted in a single system image (SSI) cluster, a system of loosely network-connected desktop computers. The main advantages of using GAs as optimization tools are for multi-objective optimization, and high probability of achieving global optimum in a complex problem. To prevent a premature convergence in the early stage of evolution for multi-objective optimization, the Pareto optimality was used as an effective criterion in offspring selections. Since the execution of the genetic algorithm (GA) in search of optimum is population-based, the computations can be performed in parallel. In the cases of uneven computational loads a simple dynamic load-balancing scheme is proposed for optimizing the parallel efficiency. It is demonstrated that the huge amount of computing demand of the GA for complex multi-objective optimization problems can be effectively dealt with by parallel computing in an SSI cluster.

Performance Evaluation of DIRECT Algorithm in Parallel Optimization for a Thermohydrodynamic Lubrication Analysis

“…Each node contains an Intel Celeron/1.7 GHz CPU and 256 MB RAM. The operating system is Linux/RedHat 9.0 and the message passing among the nodes is performed automatically by an open source single-system image (SSI) clustering software [1][2]. In this study one computer acted as a master node and the other 16 nodes were slave nodes.…”

“…Several efficient optimization methods were proposed recently for tribological studies [1][2][3][4]. For instance, the genetic algorithms (GAs) are known for their ability to conduct multiobjective optimization in parallel due to population-based search [1]. Another parallel optimization for air bearing analysis was conducted recently by using a divide-and-conquer (DAC) scheme [2].…”

An Optimizing Technique based on Genetic Algorithm for Power Management in Heterogeneous Multi-Tier Web Clusters

“…The consumption of energy was reduced by using a feedback loop. N. Wang [25] carried out the performance of parallel optimization with the help of genetic algorithm (GA) for lubrication analysis. The author used an air-bearing design as an example and he carried out the parallel computation in a single system image (SSI) cluster.…”

Multiobjective design of porous air bearing using group inching fortification method

“…Thus, to effectively solve a problem of two or more objectives is of practical importance. Some studies show the results of recent effort in developing efficient schemes (Wang and Chang, 2004;Hirani and Suh, 2005;Wang, 2005;Bhat and Barrans, 2008;Wang and Cha, 2010;Lu and Xie, 2014) for solving multiobjective optimization problems (MOOPs). To minimize the execution time many of these optimization analyses were carried out by using parallel computing or approximating the computationally intensive function with a surrogate model (Li, et al, 2008;Srirat, et al, 2012).…”

“…A few pioneer works for solving multiobjective optimization problems using evolutionary or genetic algorithms in engineering applications can be found in the literature (e.g., Deb, 2001;Coello Coello, et al, 2013;Knowles, et al, 2008;Branke, et al, 2008;Li, et al, 2008). Some recent MOOP solvers for tribological designs are adopted from modifying single-objective optimization algorithms, such as the multiobjective genetic algorithm (GA) (Wang and Chang, 2004;Hirani and Suh, 2005;Wang, 2005;Bhat and Barrans, 2008;Chiba, et al, 2014) and hyper-cube dividing method (HDM) (Wang and Cha, 2010).…”