Difference Between Social Learning And Individual Learning In A Bertrand Oligopoly With Sequential Consumer Search

Vriend (2000) proposes using two variants of a Genetic Algorithm for studying the difference between individual learning and social learning in Cournot game and obtains convergence to Cournot equilibrium with individual learning, convergence to Walrasian equilibrium with social learning. This article is to study the difference between individual learning and social learning in Bertrand oligopoly with a sequential search. We use GA for modeling the learning behavior of firms in repeated Bertrand games and use a belief-based learning for modeling the consumers'search. Find that, in this market for homogeneous product the processes of dynamic adjustment of individual learning and social learning are more complicated. In Bertrand oligopoly with individual learning, number effects play an important role. With two firms and very low search cost, equilibrium does not exist. But with social learning number of firms has no effect on market equilibrium. The market prices with social learning rise and approach to the equilibrium value of with individual learning during the early several thousand trade-days, then jump to the competitive level and keep stable at the competitive price level. In Bertrand oligopoly with individual learning, with more than two firms the market price may be any value between monopoly price and competitive price depending on search cost.