Additional Algorithm Properties

• Additional members are generated using crossover.
• Crossover takes two parent hypothesis from the current population and creates two offspring hypothesis by recombining portions of both parents. The parents are chosen probabilistically using the same formula mentioned above. Now the new generation contains the desired number of members.
• Now a certain fraction of these members are chosen at random, and random mutations are performed.
• randomized, parallel beam search for hypothesis that perform well according to the fitness function.