EA-based Problem Solving Environment Over the GRID.
Author | : Mohamed Wahib |
Publisher | : |
Total Pages | : |
Release | : 2008 |
ISBN-10 | : 9537619117 |
ISBN-13 | : 9789537619114 |
Rating | : 4/5 (114 Downloads) |
Download or read book EA-based Problem Solving Environment Over the GRID. written by Mohamed Wahib and published by . This book was released on 2008 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: This chapter presented a grid based problem solving environment that uses EAs and other algorithms all falling under the meta heuristics category to offer black box global optimization for the user. The chapter first highlighted the grid computing technology and then discussed with reasons behind using the grid for MHGrid, Meta Heuristics Grid, and the benefits of the grid technology compared to other distributed paradigms. Then a comparison of MHGrid with related work was discussed, to imply the concepts behind the design of optimization solving grid applications. The design and implementation of MHGrid was explained, including the layered architecture, the workflow inside the framework and explanation of MHAPI, a library that allows the solver developers to integrate their solvers with MHGrid. MHGrid as a model was expanded in both the vertical and horizontal directions in order to widen the base of MHGrid to be a general framework rather than being tailored to one problem type. The expansion strategies reformed the architecture of MHGrid into a SOA, the main impact for MHGrid adopting SOA was the representation of solvers and objective functions as services and thus having the service oriented grid application mostly affecting the application layer whilst using OGSA and Web services at the middleware layer. A sample example case was demonstrated to acknowledge the reader with the user perspective of MHGrid. For the future work, modifications and extensions will cover different aspects. Major points will include adopting a more sophisticated SLA mechanism, defining new interfaces that allow one solver to use another solver, for example pBOA algorithm can internally use Tabu search for candidate offspring selection, and one more important point is to conduct more study on the dynamic grain size in EAs to reach the best formulation of parallelization models adopted. Other minor points will include enchantments on the portlets to auto generate the MHML files on behalf of the users.