Progress Report

Over the last few weeks I have been progressing steadily in an attempt to finish my honours project by the set deadline. I have progressed in both the practical side and the technical writing side of the project and this report will outline this.

The progress I have made in the practical side of things, whilst hindered by small errors, has been mostly steady. So far I have created the graphical framework that the AI will be based in. During this process, I was slowed down by having to revise how the engine works and dealing with other problems that cropped up along the way. It is still in the early stages, but by having this completed so early in the term, I am now safe in the knowledge that something will appear on screen by the end of the project.

I have also made good headway on the actual AI itself. I have created a small basic application that creates decision trees. This is the ground work on what will be the final application. At the moment it is hard coded and only can only be navigated by the user. By the end of the project the application will build these trees based on the past experiences of the AI agent dynamically, reaching the most appropriate decision, for that scenario.

To further this goal, most my recent work has been put into implementing the ID3 algorithm. This algorithm uses the concept of entropy and the “knowledge” of the intelligent agent to build tree decision tree. So far this has been the most challenging aspect of the project so far and as a result I am slightly behind as to where I should be based on the time line I set myself at the start of the year. Despite this set back it is a crucial part of the program and without it none of the rest of the project will matter. However after a few extra days on solving this problem I have managed to solve the issues and am now finished creating the algorithm that will eventually lead to a decision tree being created.

The writing is also starting to take shape. At the moment, I have separated the dissertation into the sections that I believe will be the most crucial. As I progress or an issue occurs, I write short notes under the appropriate section. As of right now, I have written a draft abstract, 700 words of the introduction and around 500 words of notes on the two mentioned techniques. It is hoped that these notes will be used and the gaps filled in at the end of the project and the dissertation finished with little or no trouble.

The next step that I plan to take in the project implementation is to combine the ID3 algorithm and the decision tree code so that an AI agent can actually make decisions. This should take no longer than one week to complete as I believe I may have to make slight changes to the structure of the code I have already written in order for the two to fit together. However once this is finished the main hurdle for the project has been overcome and the project needs only be refined.

As I have already said I am now slightly behind in the project, I hope to make up a lot of time over the coming weeks after the two techniques have been merged. I set myself probably more time that absolutely necessary to complete the graphical aspect of the project so I should be able to make up time at this stage provided all things go smoothly. Please see below for an updated Gannt Chart for further information as to the changes in the time line.