Application finished

Well long time since I last wrote, here have basically been coding and working furiously since then.

Got the application more or less finished, still want to add in a proper GUI but having some issues with it at the minute. The path finding I've got is like a mix from A* and some other techniques but mostly A*. The problem is it sometimes crashes out and breaks the program, which is not ideal. I think this could be resolved but I've been struggling with it for days now and don't seem to be making much progress, any time i fix one thing, another 2 things stop working etc.

Anyways as i spoke about before have been working on the dissertation somewhat as well, and thats coming along nicely. I need to swap out some references and still add in a couple of sections here and there, but all in all it seems to be going almost to schedule. Will be a nice relief when this is all done!

Quick update

Just a quick update today! well the pathfinding is still in its conceptual stage, everything is set up ready for its implementation with the agents being controlled by keyboard input still. But when they move the map changes with their position as was seen in the screenshots earlier. I have also thought of a great way to record the results of this project and display them in a an easy artistic way.

I will store the number of moves it takes for each agent to complete traversing the terrain. This data will then hopefully be written to a separate text file for each agent and then be used to formulate a graph detailing which agent (if any) and thus which behaviour type has been most successful over time. Another feature that could be useful for research is the idea of having the application run at different speeds with the option to have it autoplay. What I mean by this is to be able to run the application so it runs much at the pace of a modern game would, so the user can switch views and see how each agent progresses. Another option could be to have the program run at its maximum possible speed and to automatically reset and re-run each time the final agent reaches its destination.

This of course would be useless from a viewers perspective, but it could potentially give rise to a vast amount of test data very quickly, if the application was left running say over night. Anyways these are just some of my quick thoughts on ways of completing this project and possible conclusions to meet for the upcomming presentation /panic again :D

Next steps

The next stage in my application is to get the A* / D* pathfinding working properly. Once this is in place i can move onto the behaviour types to get them working and then finally I can start to see some results from my research.

Also have started working away on the dissertation. According to Dr Ozveren I should have been started on this long ago, but I still feel pretty confident that I will be able to complete this within the time frame available. My literature review is pretty much finished and I have had a few drafts at various pieces of the write up so far. None worthy of posting here mind you :P

I have prepared some screenshots in the next post which detail the different mapviews. The first screen shot is the default view for the user. The entire map can be seen and the agents also can be seen on their respective tiles. The 2nd, 3rd, 4th and 5th screenshots all show the different maps stored for each agent unit (each instance of the agent class) At the bottom of each screen shot you can see the placeholder texture for the GUI :p

The pathfinding is the last big challenge ahead of me, as this task will run in parallel with the behaviour programming of the agents. Its quite exciting to start nearing the finish of a project that was conceived before Christmas. Still however much has yet to be done. On with it !

Map Screenshots

One month to go :O

Well its April now, meaning just over a month left of this my final year. It seems like its gone by in a flash, really where did all that time go ? :P

Well anyways, I have made some pretty huge upgrades to my application since the last entry. I have fully implemented my Agent class and started briefly on the GUI class. The Gui class has actually very little coding to it, and could actually be unworthy of a separate class, but it makes organisation so much easier to keep it this way. For the minute at least the class only renders a placeholder texture but the functionality is in place to render a proper GUI which will be worked on in the near future.

The agent class stores a separate map for each agent and depending on the input from the user so that there are 5 different terrains which can be displayed on screen, at any time. One thing to note is that all of the maps are the same in their overall values, but its just that only the sections that have been discovered by the individual agents that are displayed.

New Map Screenshot!

Well my map class is finally starting to come together, Ive edited the positioning somewhat so its more aesthetically pleasing and easier to see whats going on. The proper textures are now in place so that I can see what the map is really going to look like. I've edited the size some, to a size of 80x40 squares. This is to make space at the bottom - the black section that is going to be for the GUI to relay information to the user.

The next goals to work on will be the assignment of the starting and ending zones on the map, and maybe play about some with the density of each tile texture and value, to make maybe more mountains and less desert for example.

Once this is completed, it will be time for construction of the Agent class and the GUI class. What I want is for each agent to have their own version of their map which the user can select and display at any time, meaning the agents progress can be tracked in real time, with the user able to see their known and unknown terrain and how they each individually react throughout their respective journeys.

Current Progress

Well since my last post, i have met with various tutors including Dr King and Dr Lucas about my absence the last couple of months. Also since then I have not really made much progress with my map generator, but instead I have focused on making my code fully object oriented rather than the single C file that it was.

So far I've managed to separate out the code so that there is improved flexibility without losing any of the functionality. As well as a separate map class, I now also have classes for the loading of bitmaps and the conversion of these bitmaps into textures for use within my application. Now that I have this framework in motion, the next step will be to finish off with my map generator class to move onto the next step which will be the implementation of the agent class.