Michael LaLena

Michael C. LaLena is a computer scientist working for Carestream Health where he designs software for computed radiography and digital radiography systems.

LaLena received his Master's Degree in computer science from Rochester Institute of Technology in 1997. Since receiving his degree, he continues to work in the Artificial Life field investigating and experimenting with new technologies.

LaLena was the first to investigate “required” teamwork between actors in Genetic Programming – specifically the ant colony food collection problem. Previous work with swarm intelligence or ant colony optimization has ants communicating with pheromones or moving in loose knit groups. While this communication allows a large number of ants to solve the problem quickly, these ants are not required to solve the problem. One ant could just as easily solve the problem given enough time, making the communication between ants not completely necessary.

Food collection problems involving required teamwork require more than one ant to solve the problem. One example would be collecting food that requires more than one ant to lift. Multiple ants must work together to pick up the food and return it to the nest. This problem becomes very difficult if there is more food than ants. For example: If there are 10 pieces of food each requiring 10 ants to lift and there are only 20 ants available to lift the food, then teamwork is required to solve the problem. If each ant randomly searched for food and stopped when it found food, you might have a situation where every ant is waiting for help from other ants and no more ants are available.

Another example is a water crossing problem where ants must sacrifice themselves to allow other ants to cross a river to reach food. If every ant sacrificed himself, then there wouldn't be any ants left to collect the food. If no ant is willing to sacrifice himself, then the ants will not be able to get to the food. This problem requires a special type of teamwork where some ants must sacrifice themselves, and others must collect food even though every ant is running the same program and there are no state variables to assign different duties to each ant.

This research has practical swarm robotics applications for military robots in combat and support roles and in exploration.

Main Menu

See Also

Main Menu