|
::: HELP WANTED!!! :::
There is a number of mobile service projects I would like to developed using 3APL-M but I just do not have the time. I know there are students out there looking for ideas on how to create a agent-based, java coded, academic appealing, mobile device deployable application for a end of term project, demonstration or just for fun. Hence, why not join efforts and
collaborate
?! I am presenting some ideas for projects, issues to be fixed and support applications.
If you are interest or you have ideas for other projects or any other reference material please drop me an email into
fkoch at acm dot org
.
Thanks!!!!
FOR MORE EXAMPLES...
This is a rather static page and present some selected examples only. For more examples, check my personal HELP WANTED!!! page
here
.
HELP WITH Performance tests
click to enlarge
click to enlarge
I would like to run a set of performance tests on the 3APL-M deliberation machine in several platforms and (1) check how the deliberation process performs for in different computing environments and; (2) test the performance per knowledge loaded.
For (2), my idea is to construct some load test applications that load the 3APL-M knowledge information with different set of information (like dozens of plans to chose from, plans with several levels of sub-planning and so on) and measure the performance. I would take the measures using a range of values, put in a spreadsheet, draw the graphs and analyse the values. From there it would be possible to get to the conclusion on what kind of problems a 3APL-M program can solve (depending on the number of plan rules, belief base entries and depth of plan resolution). Then, for (1) I could make the same analysis in different hardware scenarios (PDA, phones, handhelds, etc...).
I have done some preliminary work, but that would have to be extended!
PacTest001.java
PacTest002.java
PlanRuleBaseTree.java
example of incomplete report (idea!)
HELP WITH Creating an Intelligent Assistant
![Mobile device support for mobile users (from [Rahwan2004])](help/abima.jpg)
click to enlarge
I would like to convert the work from [Rahwan2004] (see reference below), who created the Agent-based Intelligent Mobile Assistant (AbIMA) in AgentSpeak(L) into 3APL-M. "AbIMA will use agent-based techniques to reason about these mental attitudes and make decisions about what plans (and subsequently, what actions) to suggest to the user. Another attractive notion of BDI architectures is that they are usually based on the aggregation of plans from plan libraries, encoded for a particular domain. This results in significant reduction of the complexities encountered in traditional planning systems, which is particularly helpful within the context of operation on limited mobile devices. Hence AbIMA will have a repository of pre-programmed plans that can be used and combined in order to achieve different goals. In particular, AbIMA is based on a specific BDI architecture and programming language called AgentSpeak(L)"
T. Rahwan, T. Rahwan, I. Rahwan and R. Ashri
Agent-based Support for Mobile Users using AgentSpeak(L)
Agent Oriented Information Systems, LNAI, Springer Verlag, 2004.
HELP WITH Creating an Intelligent Assistant
click to enlarge
I would like to develop a Personal Assistant demo involving location-based services, (light) personal asstance and negotiation. For example, in the scenarion at the Figure Layla is at BUILDING B and wants to go home. Her agent sends a query to a list of nearby friends asking for a lift. Omar, at BUILDING A replies saying he is going home around the same time Layla wants. But as Omar needs to pickup a book at BUILDING B, his agent counter-proposes to Layla, asking if she could do him this favor. Layla accepts and also requests for Omar's car position, so they can meet there. All these negotiation steps are done by agents, following the negotiation rules (maybe Interest-Based Negotiation as in [Rahwan2003]).
I. Rahwan, L. Sonenberg and F. Dignum
On Interest-Based Negotiation
Advances in Agent Communication, LNAI 2922, Springer Verlag, 2003.
|
