Multi Agent Path Finding

A multiagent system I implemented to find the shortest route through this maze. It is modelled on the way an ant colony locates food. Each agent (the yellow dots) can only see the squares immediately above, below, to the left and to the right of its current position. It selects which square to move to at random. The likelihood of each square being selected is weighted based on the number of agents that reached the end that had passed over that square, how long they took to get there and how long