Hello Judges!

Our robot was coded entirely in Java. Our strategy relies heavily on
so-called "Points of Interest". Clearly, given the limited information
available to us throughout the game, certain decisions had to be made
regarding how important we consider different aspects of the game and
how much these aspects should figure into our strategy. In the end,
we decided on several POIs (ex: Home Bases, the destinations of all
packages that we're carrying, etc.) and then, had a fitness function
for each of these POIs that would assign it a score, based on certain
factors (ex: How far we are from the POI, how many points we get for
delivering this POI (assuming it's a package), "hunting" mode vs. "delivering" mode, etc.)

Next, the coefficients for the formulae used in the fitness functions
were determined using a genetic algorithm. We played off 10 robots, and
then kept the 4 that obtained the highest score, mutated some of their
genes and produced another 6 and repeated this process. In this way,
by tweaking the coefficients properly, we would always tend to the most
beneficial POI. Of course, this is glossing over many of the details of the project, but our hope is that this somehwat explains the thoughts behind our robot.

Team MonkeyJuicers