An agent-based approch to grid programming

Computational grids have provided the usage of computational distributed resources for computation-intensive applications. The development of programs, the brain, that use these capabilities, the brawn, is one of the challenging issues for grid computing. In this paper, an effort has been made in order to tackle this problem by applying a mobile agent-based programming model on the grid. This model called Alchemi+ has been materialized by extending Alchemi (TM) grid infrastructure and adding agent properties and navigational commands. Alchemi+ will let the user develop his parallel program using agents' mobility and communication between them. In order to evaluate the system, a matrix multiplication algorithm as well as an algorithm for finding the convex hull of a series of points have been implemented in the mentioned system.