Using Pareto Front for a Consensus Building, Human Based, Genetic Algorithm


Pietro Speroni di Fenizio, Chris Anderson,
Using Pareto Front for a Consensus Building, Human Based, Genetic Algorithm
Proc. European Conference on Artificial Life (ECAL 2009), LNCS, Springer, Berlin, pg 175-182, 2009
Preprint Poster

Abstract:

We present a decision-making procedure, for a problem where no solution is known a priori. The decision-making procedure is a human powered genetic algorithm that uses human beings to produce variations and evaluate the partial solution proposed. Following [1] we then select the pareto front of the proposed partial solutions, eliminating the dominated ones. We then feed the results back to the human beings, asking them to find alternative proposals that integrate and synthesise the solutions in the pareto front. The algorithm is currently being implemented, and some preliminary results are presented here. Some possible variations on the algorithm, and some of its limitations, are also discussed.

After presenting this work in Budapest, I was invited to present it again in Southampton. And then again in Coimbra. A video mashed up from those talks is available.