Lab 11: The strategy pattern

The goal of this lab is to practive using the strategy pattern.

1. Introduction and setup

Today we go back to the beloved predator/prey simulation. This time we will be adjusting the algorithm the animals use to find prey and elude predators.

The current version of the simulation can be copied from a subdirectory of the course public directory.

cp /homeemp/tvandrun/pub/245/simulation3/* .

The algorithm that all animals use is to scan the the visible area around it, starting with the upper left hand corner, and given the first thing seen that is either a predator or a prey,

• If it is a predator, move in the opposite direction as far as possible.
• If it is a prey, move on it; or, if it is out of range, move towards it as far as possible.

We have seen the effects of this. Rabbits congregate on the upper left can corner of a patch of clover-- either moving right into a predator's mouth even though a there's a predator-free clover-filled area nearby, or forming a swarm which moves from left to right. Animals lower on the food chain get away from predators above and to the left of them, but are practically blind to those below or to the right. Predators chase rabbits to the left of the board, or compete with each other for the same rabbit, when a swarm is only a few squares to the right. These animals could help their survival and the survival of their species if they only had smarter instincts.

In the current version of the simulation, the behavior of the scan method is mostly delegated to a strategy. Look at the code to familiarize yourself with how it works.

2. Writing new strategies

Your task is to encode new behavior for the animals by writing new ScanStrategy classes. Some things to think about:

• The most obvious changes are for an animal to go after the closest prey, and to make eluding predators a higher priority than catching prey.
• Since weight() is a public method, a predator can pick its prey based on which is the fattest.
• One handicap the old algorithm puts on predators is that they can't remember what animal they are chasing. They often "get distracted." The StupidStrategy is a singleton, but Strategies don't have to be. To remember who you're chasing, the strategy might need state.
• In real life, locusts and bees behave differently when they are swarming. Not that rabbits need any help in this matter (once a rabbit "swarm" has started, there is no stopping it), but it might make sense for rabbits to become bolder if they see a lot of other rabbits around.
• Some strategies might be better for animals higher or lower on the food chain, with better or worse visual acuity, etc.

Experiment. Start out by observing the current behavior. There basically are two possible outcomes: The clover wins, because all the rabbits are eaten before the clover is eaten, and then the clover expands with nothing to stop it; or, the rabbits win, by forming an unstoppable swarm that sweeps the clover off the map. I have found that if we hold constant the number of clover at 100, rabbits at 20, cougars at 1, and bears at 1 and vary the number of foxes, the boundary is somewhere around 12 foxes. At 10 foxes, the rabbits almost always win. At 15 foxes, the clover almost always wins.