"swarming.c" simulates the swarming of a school of fish in a light field. Reinforcement Learning is used to train the school of fish to optimize the swarming behavior. The resulting performance is acheived by finidng the optimized values of the underlying interaction parameters. The final aim is to see if several different fractions of the population having different interaction leads to any interesting segregation phenomena. "reward_landscape.c" generates the averaged reward function in the parameter space. "filed.nb" is a Mathematica script that evaluates the maximum and the minium of the field that are used to normalize the field. "max_min02.dat" contains the maximum and minimum values of the different realizations of the field.