Well been a couple of weeks so here is were I am now:
I have a rough start on my writeup started, I have almost a 1000 words on the "Dynamic Scripting" technique but not too happy with them so I will be re-wording them some time next week.
However the big news is that the application is almost finished now, I haven't been able to implement the actual gathering of rules into the LUA script however I have successful write a collection of 3 C++ programs that:
1. Sort The State Rules Bases into descending order based upon the weights values of the rules in that particular state.
2. A program that then extracts the rules from these State Files and compiles a script for implementation in WARGUS, this file does as it should and moves onto the next state file after a state change rule is called.
and 3. I am halfway through the program that will adjust the weights values of all the rules in the files after a match has been concluded.
This means that hopefully the application will be finished by the beginning of next week and testing will only take a day or so and so I should have a few weeks to complete the writeup stage of the project!
Thursday, 26 March 2009
Tuesday, 10 March 2009
Progress Reporting
Weekly Progress Report
Stuart Robb’s Honours Project
Week: 25 ( 9/03/09 – 15/03/09 ) Weeks Left: 10
What Was Going To Be Done For This Week:
This is the first progress report and therefore no formal goal as such was set for this week, as the original plan (Gantt Chart) has become obsolete as the project moved into the new year.
What Was Done This Week:
This past week has seen a small road block towards moving the project forwards for two reasons:
The main reason was as a one off class test was to be held this week much time was spent revising for that as the project was at a comfortable stage that light reading would suffice for the week.
The other reason for a lack of progress was my desktop PC broke once and then after being repaired broke again; this lead to a large amount of time being distracted from all forms of work as the problem if not fixed rendered the computer useless.
So what was actually completed this week towards project goals was the technology tree for the application game was calculated, this was necessary before any progress could be made with implementing Dynamic Scripting into the application since the rules base is dependent on the different states in the game which are defined by the technology tree.
Overall Project Progress:
The project is now at a point where the majority of the reading about Dynamic Scripting has been completed and this frees up reading time to be spent on other parts of the literature review such as background into reinforcement learning as well as a paper on how real time strategy games are used as the benchmark for testing high quality AI.
The application for the tests for this project is also quite far along as the installer found on the internet allows for the easy implementation of a custom AI script into the game. With the technology tree completed it is now just a matter of compiling the initial rules base for the AI and then implementing a custom script that can be altered between matches, this is properly going to be the hardest part of the implementation but also the final part.
No attempt at beginning the writing of the dissertation has been started yet although the plan is to start moving forward with the literature review section covering Dynamic Scripting. Also the bibliography is being constantly kept up to date as new articles and books are read.
What Is To Be Done For Next Week:
This week’s objectives are to get at least 1000 words done for the literature review section on Dynamic Scripting, have a draft copy of the initial rules base for the AI which can then be optimised and to keep up reading on reinforcement learning.
Stuart Robb’s Honours Project
Week: 25 ( 9/03/09 – 15/03/09 ) Weeks Left: 10
What Was Going To Be Done For This Week:
This is the first progress report and therefore no formal goal as such was set for this week, as the original plan (Gantt Chart) has become obsolete as the project moved into the new year.
What Was Done This Week:
This past week has seen a small road block towards moving the project forwards for two reasons:
The main reason was as a one off class test was to be held this week much time was spent revising for that as the project was at a comfortable stage that light reading would suffice for the week.
The other reason for a lack of progress was my desktop PC broke once and then after being repaired broke again; this lead to a large amount of time being distracted from all forms of work as the problem if not fixed rendered the computer useless.
So what was actually completed this week towards project goals was the technology tree for the application game was calculated, this was necessary before any progress could be made with implementing Dynamic Scripting into the application since the rules base is dependent on the different states in the game which are defined by the technology tree.
Overall Project Progress:
The project is now at a point where the majority of the reading about Dynamic Scripting has been completed and this frees up reading time to be spent on other parts of the literature review such as background into reinforcement learning as well as a paper on how real time strategy games are used as the benchmark for testing high quality AI.
The application for the tests for this project is also quite far along as the installer found on the internet allows for the easy implementation of a custom AI script into the game. With the technology tree completed it is now just a matter of compiling the initial rules base for the AI and then implementing a custom script that can be altered between matches, this is properly going to be the hardest part of the implementation but also the final part.
No attempt at beginning the writing of the dissertation has been started yet although the plan is to start moving forward with the literature review section covering Dynamic Scripting. Also the bibliography is being constantly kept up to date as new articles and books are read.
What Is To Be Done For Next Week:
This week’s objectives are to get at least 1000 words done for the literature review section on Dynamic Scripting, have a draft copy of the initial rules base for the AI which can then be optimised and to keep up reading on reinforcement learning.
Subscribe to:
Comments (Atom)