It’s been a very long 11-day weekend, with many ups and downs. To compensate for having no lectures for a good 11 days, we’d been given Prolog coursework involving monitoring behaviour of swarm AI in multi-agent systems. It involves a simulated environment called TokensWorld where several species of agents exist. The agents which inhabit TokensWorld are either staring into a void (pile of 0 tokens) or standing in front of a pile of n tokens (where in our version of the world, maximum n = 8). Agents can wander to another pile (generated by random), grab or drop a token into a pile. The interesting part arose when we realised the goals of the different species.
The Kleptomaniac Agent. Its goal is to collect all 8 tokens into one pile and be facing it, as opposed to the Vandal Agent, whose goal is to break down any attempt at building by distributing the tokens into 8 piles of 1 token each. A third agent, the Voyeur, is content with the goal of merely observing a pile of 3 tokens. One of our tasks was to help the Kleptomaniac achieve its goal more often (increase its success rate over 10,000/100,000 simulated runs). Another task was to introduce a new agent species called the Artist, whose goals involved building two piles of 1/3 tokens each, and observing such a pile. Allowing multiple instances of different species to interact in simulations produced interesting results and helped determine which species had easier lives in this world.
“Life for the artist is even tougher than for the kleptomaniac”
That was the easy part. We’ve now descended into generating Discrete Situation/Policy graphs and solving Bellman Equations which would algebraically deduce policies to help species achieve their goals. The coursework is followed by a Prolog test then a short course in the C programming language.
In other news, my main laptop, the HP HDX-18 1010ea has died by way of a hard drive failure. It failed roughly two days ago with the only warning being an Explorer crash while writing a program in Haskell. Restarting gave frequent blue screens and failure to boot, even in safe mode. Slightly alarmed as to why such a powerful and seemingly healthy machine would give up like this, I started up the HP Pavilion TX1250-ea which added to the frustration by having a glitchy display. I then resorted to my mid-range Gateway machine to fix the second laptop and began an effort of diagnosing the problem on the first. My initial priority was to launch an incremental backup to my external hard drive due to the lack of committing a change since late April, raising potential to lose over 900 unprocessed images from recent events.
Luckily, I was able to perform this incremental backup overnight as I slept to the lullaby of hard disk crackling sounds. The next day, with a full backup, I aggressively probed the system to try to deduce the problem further, and was surprisingly able to install Windows 7 and restore it to working order. This, unfortunately, did not last. Although still unsure what the problem really is, I am relatively certain it remains a hardware fault with the disk at logical position 0. The laptop failed again with disk I/O complaints and failure to boot, resulting in the unfortunate waste of a Windows 7 Professional x64 key. I will send it to HP Support to get the disk replaced.
Until then, I have my third laptop hooked up to my monitor in what I can only describe as a chaotic bid to salvage a working terminal for general use with a flurry of wires everywhere. The entropy and technological disorder of my room is through the roof, so I’ve retreated back home for some peace and stability for a short while. I shall return to reorganise the setup of my terminal and attempt to restore order in a few days. The 11 day weekend continues until this Wednesday, when the Prolog test is held. Although I remain relatively stoic about these events, I do hope things start looking up soon.