The Research Of The Venture Capital Equity Exit Auction Based On The Multiple-agents Simulation

Venture capital can promote technological innovation and economicdevelopment, and there are four main ways in venture capital exit: the initial publicoffering, mergers and acquisitions, repurchase and liquidation. The threshold ofparticipating in Growth Enterprise Market is relatively high in China, thus initialpublic offering is accordingly expensive; merger and acquisition generally exitinformation asymmetry, leading to a undervalued valuation of risk enterprise orproject, making the true value of the venture capital project can not be effectivelyrevealed, so venture capital is difficult to successfully and efficiently exit the market;due to risk enterprise’s funding problems, difficulty of repurchases is also very high;liquidation is only adopted in corporate bankruptcy. Although these methods havetheir own advantages, but can not achieve maximization of venture capitalists’ capitalvalue. Solving the core of the problem, we should reveal the true value of the venturecapital project in the condition of information asymmetry, prompt the capital ofinvestors realizing capital value maximization, finally achieving the purpose ofventure capital appreciation and recycling.Auction theory solves the problems of maximizing the value of goods andresource allocation optimization, because the auction is essentially an effective meansof allocation of resources in the case of an asymmetric information and valuerevealment mechanism, and thus successfully avoiding moral hazard and adverseselection due to asymmetric information in the exit of venture capital. This articlestands in the seller’s perspective, establishing two kinds of multi-attribute auctionmechanism mathematical model, namely the uniform price auction model anddiscriminatory pricing model, for the purpose of maximizing venture capital equityvalue. There are a variety of quality attributes of the venture capital equity such asprice, delivery time, bidders’ reputation, payment terms, value-added services and soon, this paper uses multi-utility theory, broaching a two-stage auction system toscreening bidders, in the first stage applying multi-utility theory to obtain comprehensive scoring, thus roguing some low quality bidders; in the second stageswitch to the most important price attribute, similar to a single attribute items auction,at the same time dividing the price auction into a uniform price auction anddiscriminatory price auction, respectively, establishing the correspondingmathematical model.The auction process involves the behavior of bidders, aimed at characteristics ofthe associated value auction of venture capital equity exit, such as bidders’ intelligentbehavior, co-learning, information transmission, as well as strategies adopted by theauction bidders, this article using multi-Agent systems to introduce auctionmechanism to solve the problem of equity value maximization and resource allocation.We assume that these intelligent participants (Agent) is the bidders, in an open、dynamic multi-Agent systems, self-interested Agent has its own goals and preferences,through observation, consultation and deduction to determine what kind of offerbehavioral strategies to be adopted, in most cases single Agent can’t achieve targetsindependently, needs the participation and help of other Agents, in other words,multiple Agents should jointly solve the offer optimization problem throughcooperation. According to the revelation principle of the auction, when bidders offerto tell the true valuation, it’s their own optimal strategy, the auctioneer can alsoachieve the greatest profits, the auction system will reach an equilibrium, but incondition of associated value model, it’s difficult to make bidders tell the truth, as thehardness of excluding all the noise in the market. so using learning optimizationmechanism in intelligent algorithm to allow bidders achieve the optimal solutionalong with the intelligent groups’ optimization, reach the balance of auction system,so that obtain the greatest expected profits of venture capital exit equity auction. Inthis paper, particle swarm optimization (PSO) is applied to optimize bidders’ biddingstrategies in the condition of unified price auction mathematical model, and geneticalgorithms (GA) is applied in the condition of discriminatory price auctionmathematical model.The associated value auction model and the bidder’s bidding strategy can be seenas a dynamic complex optimization system, traditional mathematical and experimental methods is difficult to seek the specific solution, this paper usingresearch methods of complex systems-multi-Agent simulation to solve complexauction bidding optimization and expected revenue maximization problem.Multi-Agent modeling and simulation is a very flexible technology, its essentialfeature is to establish a conceptual model from the perspective of multi-Agent. Thearticle is based on multi-Agent perspective to describe venture capital exit equityauction, in the process of auction bidding, bidders are Agents in the associated valueauctions, the thinking and behavior of these reality-based bidders are complex, theyhave the features of intelligence, adaptability, learning and interaction, among Agentsnoise inevitably exits in the information communication channels, Agents can obtainpart of the bidding information from others consciously or unconsciously to realizeheir own interests maximization. The above-mentioned complex selection processmeets the characteristics and properties of multi-Agent, so this article choosesmulti-Agent simulation platform netlogo to simulate bidders’ bidding strategies andthe uniform price auction model and discriminatory price auction model that havebeen established.The auction process and the main process in the model can beconcluded as actual system�'conceptual model�'simulation model�'simulationresults�', conclusion and finally reaches a verification and correction of venturecapital equity auction models’ correctness and rationality.