Strategic Learning in Multiple Equilibria for Double Bargaining Mechanism by PSO
The learning behaviours of buyers and sellers with the assumption of bounded rationality were studied in the double sealed-bid bargaining mechanism. A multi-agent simulation trading system was constructed to observe the process of equilibrium approach when exist the multiple equilibria. The bidding choices of the agents were modelled by particle swarm optimization (PSO) algorithm. In our proposed model, two populations of buyers and sellers were randomly matched to deal repeatedly until the iteration stop, and each agent would update his bidding strategy in each round by imitating the successful member in his population and by private experience. Results show that the final biddings of the agents in both populations commonly approach a Nash equilibrium which is reasonable for the market principle.