Abstract:
This paper elaborates the use of distributed Genetic Algorithms (DGA) to study an artificial land rental market. The
study is based on a spatial comparative-static model in which a number of spatially ordered agents (farms) compete in an auction
for renting land. Each agent’s behavior is determined by a genetic algorithm that is applied to an agent specific population
of genomes representing particular bidding strategies. Agents interact directly through a migration mechanism that allows to
spread renting strategies across the population of agents as well as indirectly over the rental market. Two market constellations
are considered and different simulations with a variety of parameter constellations (migration rate, placement of farms,
etc.) are run: First, a situation of limited market access is defined. A series of simulation experiments shows that for this scenario
the DGA generates results that fit comparative static equilibrium conditions like allocative efficiency and zero-profits.
Second, in a limited market access scenario, only under very special conditions the DGA generates results that comply with
oligopolistic behavior. The results of the two scenarios are analyzed and discussed as to the influence of the DGA procedure
itself and a possible economic and game theoretic interpretation.
