A Trade Network Game with Endogenous Partner Selection

This paper develops an evolutionary trade network game (TNG) that combines evolutionary game play with endogenous partner selection. Successive generations of resource-constrained buyers and sellers choose and refuse trade partners on the basis of continually updated expected payoffs. Trade partner selection takes place in accordance with a modified Gale-Shapley matching mechanism, and trades are implemented using trade strategies evolved via a standardly specified genetic algorithm. The trade partnerships resulting from the matching mechanism are shown to be core stable and Pareto optimal in each successive trade cycle. Nevertheless, computer experiments suggest that these static optimality properties may be inadequate measures of optimality from an evolutionary perspective. Related work can be accessed at: http://www.econ.iastate.edu/tesfatsi/tnghome.htm