Research On The Optimal Incentive Mechanism Of Crowdsourcing Contest Under Different Way Of Winner-determination

In order to help companies cope with the complex, subtle and rapidly changing market demand, open innovation has been widely applied in various kinds of economic and social problems. All aspects of crowdsourcing contest have been researched by the domestic and foreign scholars in detail. An important issue of crowdsourcing contest is that how to design the incentive mechanism to improve innovation performance. Although, a few literatures use the game theory and experimental method to study the problem of incentive mechanism and reward in traditional competition, there is still a lot of conflicting research results. And few researchers take account of the characteristics of crowdsourcing contest. Therefore, this paper attempts to use the principal-agent theory to investigate the problem of the optimal incentive mechanism under different way of winner determination, and provide guidance for companies to use crowdsourcing contest effectively.Based on the prior research results, this paper firstly builds the model of winner-takes-all crowdsourcing contest. We use the principal-agent theory to analyze the incentive compatibility between organizer and solvers in different incentive mechanism. In this part, we focus on the effect of solver's participating experience on solver's effort and organizer's expected return. Secondly, we analyze the incentive compatibility between organizer and solver under the circumstance of multiple prizes. In this part, we focus on the interaction effect between solvers on solver's effort and organizer's expected return. Lastly, we verify the theoretical results through dynamic simulation. Some conclusions are given as followed.(1) No matter the case of winner-takes-all or multiple prize, there is no the incentive compatibility result in fixed prize mechanism. But we can get the incentive compatibility results in reward auction mechanism. On the other hand, in order to maximize organizer's expected return, fixed prize is superior to reward auction, when we set a moderate range of0 w.(2) The solver's participating experience has positive and negative aspects of effect on organizer's expected revenue. As solver's participating experience growing, the level of solver's effort grows. But, solver's participating experience can weaken the quality of innovation solution. On the whole, solver's participating experience is beneficial to improving organizer's expected return.(3) There is interaction effect between solvers in multiple prizes crowdsourcing contest, it can enhance solver's effort and then improve organizer's expected return. As the interaction effect enhancing, multiple prizes crowdsourcing contest is superior to winner-takes-all crowdsourcing contest absolutely.(4) Due to asymmetric information, the agency cost is inevitable. There are four important part of the agency cost: the differences of the quality of submission, solver's absolute risk aversion, solver's effort cost coefficient and solver's participating experience. We can improve the profit potential by reducing the agency cost.