Final Activity Report Summary - E-CAP (E-Commerce Agent Platform)
The aim of the project was development of a complete e-commerce system utilising software agents and ontologies. The key goal was the design and actual implementation of the system using an existing agent platform (JADE), and this goal has been achieved. The development process followed the tenets of software engineering. First, the requirements analysis was conducted. It was followed by UML-based formalisation of the system. Starting from the 'use case' diagram, followed by activity and sequence diagrams describing dynamical processes in the system and completed by state chart diagrams of all agents in the system. Next, the JADE-based implementation of the system ensued. During the implementation phase a number of technical issues have been solved - including the design of the data mart that is to be used for storing data needed for various adaptive processes that are to take place in the system, efficient implementation of the blackboard, ontologies and management of time. Separately ontologies of products, as well as utilisation of ontologies in the system have been studied and developed and incorporated into the system. The final system prototype has been committed to the Sourceforge repository of open source software.
Results obtained in the project have been published in 16 publications (actually published and to appear). Additional results in spin-off projects have been published in 24 publications (actually published and to appear). Further publications (at least 5 more) based on obtained results will be prepared by the end of 2008.
Results obtained in the project and its spin-off counterparts have been utilised in 44 presentations in Europe, United States and South Korea). Further presentations (at least 6 more) based on obtained results will take place by the end of 2008.
In order to fully utilise the third and fourth advantages, the TDD-CDMA system needs to be carefully designed by taking the following factors into account: (1) multi-user interference, (2) multipath propagation, (3) the rate of variance of the channel, (4) allocation of uplink and downlink slots, (5) the amount and placement of pilot symbols. One needs to consider the combined effect of all these factors.
In this project, the combined effect of all these factors is investigated, and Shannon capacity results are obtained. These capacity results are for practical systems which employ linear multiuser detectors for the uplink, and linear precoders in the downlink. Despite the amount of research and practice on CDMA systems, optimal design of training signals for these systems is still largely unknown. With the research conducted in this project, we shed light on the achievable capacity of training-aided TDD-CDMA systems, and provide insights for the design of such systems. The results obtained have the potential to improve the bandwidth efficiency of future cellular systems.
Results obtained in the project have been published in 16 publications (actually published and to appear). Additional results in spin-off projects have been published in 24 publications (actually published and to appear). Further publications (at least 5 more) based on obtained results will be prepared by the end of 2008.
Results obtained in the project and its spin-off counterparts have been utilised in 44 presentations in Europe, United States and South Korea). Further presentations (at least 6 more) based on obtained results will take place by the end of 2008.
In order to fully utilise the third and fourth advantages, the TDD-CDMA system needs to be carefully designed by taking the following factors into account: (1) multi-user interference, (2) multipath propagation, (3) the rate of variance of the channel, (4) allocation of uplink and downlink slots, (5) the amount and placement of pilot symbols. One needs to consider the combined effect of all these factors.
In this project, the combined effect of all these factors is investigated, and Shannon capacity results are obtained. These capacity results are for practical systems which employ linear multiuser detectors for the uplink, and linear precoders in the downlink. Despite the amount of research and practice on CDMA systems, optimal design of training signals for these systems is still largely unknown. With the research conducted in this project, we shed light on the achievable capacity of training-aided TDD-CDMA systems, and provide insights for the design of such systems. The results obtained have the potential to improve the bandwidth efficiency of future cellular systems.