A dynamic e-commerce model using multi agent systems
Abstract
While there has been significant growth of e-commerce, most of the tasks remam manual.
Placement of buy and/or sell requests have to be done manually from time to time. This
inconvenience occasioned by the manual processes should be addressed by current developers.
This study has addressed the need to have dynamism implemented in our e-commerce setup and
identifies multi agent systems as a possible option that can be embraced. While there have been
attempts at automating e-commerce most of the solutions are implemented conventional
programming including object oriented development which cannot fully capture human
intelligence in independent units and/or functionality. In this project, a model is designed and
implemented; the same is tested in a semi live environment to validate the usability in a
production environment. Independent units, agents interact with the environment (to capture with
the circumstances) and interact with one another to make decisions as per the defined belief
system , different e-commerce tasks are assigned to agents and these agents on the basis of the
environment would be able to make decision and communicate with one another to accomplish
the task at hand. The Multi agent implemented provides a truly scalable option for automating ecommerce.
The functionalities can easily be extended through the definition of new agents.
Publisher
School of Computing and Informatics, University of Nairobi
Description
MSc