基于多主体仿真的风险投资股权退出拍卖研究

发布时间：2018-07-14 21:14

【摘要】：风险投资能促进技术创新和经济发展，风险投资退出主要有四种方式：首次公开发行、兼并与收购、回购和清算，我国创业板上市门槛较高，首次公开发行费用昂贵；由于信息不对称普遍存在，企业兼并或收购的方式，将会导致风险企业或项目的价值被低估，从而使得风险投资项目的真实价值无法得到有效揭示，风险资本难以顺利退出，市场效率低下；由于风险企业资金问题，回购的也存在难度；清算只有是在风险企业破产时才选用的手段。以上这些退出方式虽各有优点但均难以实现风险投资者的资本价值最大化，解决该问题的核心在于揭示不对称信息下风险投资退出交易中风险企业或项目的真实价值，促使投资到风险企业的资本实现价值最大，风险投资顺利退出，达到风险投资的资本增值与循环使用的目的。因为拍卖本质上是一种非对称信息情况下配置资源的有效手段和价值揭示机制，拍卖理论能够解决物品的价值最大化和资源配置最优化题，进而规避风险投资退出时由于不对称信息所导致的道德风险和逆向选择问题。本文站在拍卖方的视角，建立了两种风险投资股权退出的多属性拍卖机制数学模型——统一价格拍卖模型和歧视价格拍卖模型，使风险投资股权价值实现最大化。在建立拍卖模型之前，由于风险投资退出股权存在诸如价格、交割时间、竞标人声誉、支付条件、增值服务等多种属性，本文运用多效用理论，分两个阶段筛选拍卖系统竞标人，在第一阶段运用多效用理论进行综合打分淘汰一些低质量投标人，在第二阶段就切换到最重要的价格属性上按单属性物品进行拍卖，并按照风险投资股权退出拍卖的最重要属性——价格属性将拍卖分为统一价格拍卖和歧视性价格拍卖，在此基础上建立相应的数学模型。针对风险投资股权退出关联价值拍卖的特性，如投标人的智能行为、相互学习、信息传递等一系列行为，以及投标人竞价策略，本文多主体系统引入拍卖机制来解决股权拍卖价值最大化和资源配置问题。假定这些智能的主体（Agent）就是买方投标人,在开放、动态的多主体系统中，自利主体具有自身的目标和偏好,通过观察、协商与推理，来确定采取什么样的报价行为策略。单主体往往不能独立完成目标，需要其它主体的参与和帮助，即需要通过多个主体合力来共同解决优化报价问题。根据拍卖理论的显示原理，当竞标者采取说真话的策略时候，对自己是最优策略，对拍卖者也是期望收益最大，拍卖系统将达到一个均衡。但是在风险投资股权退出关联价值拍卖情况下，要使具有有限理性的竞标者在排除任何噪音干扰情况下说真话很难，于是运用智能算法的学习优化机制可以使竞标者在不断随着智能群体得优化达到最优解，拍卖系统达到均衡，从而使风险投资股权退出的拍卖者期望收益最大。本文运用粒子群算法（PSO）优化风险投资股权退出时统一价格拍卖数学模型的竞标者报价策略，运用遗传算法(GA)优化风险投资股权退出时歧视性价格拍卖模型的竞标者报价策略。基于关联价值拍卖系统的模型和竞标人的报价策略可以看作是一个动态的优化复杂系统，传统的数学方法和实验方法很难对其求解，本文运用复杂系统的方法——多主体仿真方法解决复杂拍卖系统的报价优化和拍卖者期望最大化问题。多主体建模仿真是一种非常灵活的建模仿真技术，本质特征是采用多主体视角建立系统的概念模型。本文正是基于多主体视角描述风险投资股权退出拍卖，，在拍卖竞标过程中，各投标人是基于关联价值的拍卖竞标，由于竞标前各主体之间即现实中的竞标人的思想和行为具有复杂性，他们有智能性、适应性、学习性和交互性，主体之间必然有噪音形式的信息沟通渠道，主体间会获取部分他人投标信息，以争取自己利益最大化，这种复杂的选择过程满足多主体的特征和属性。因此，本文选用多主体仿真平台Netlogo对各主体报价策略和股权退出统一价格拍卖模型以及歧视价格拍卖模型进行仿真验证。本文的拍卖系统主要流程是实际系统→概念模型→仿真模型→仿真结果→结论，最后用结论来验证和修正风险投资股权退出拍卖模型的正确性和合理性。
[Abstract]:Venture capital can promote technological innovation and economic development. There are four main ways of venture capital withdrawal: first open issue, merger and acquisition, repurchase and liquidation. The threshold of the gem is high, and the initial public issue is expensive. Because of the widespread existence of information asymmetry, the enterprise merger or acquisition will lead to the risk enterprises. Or the value of the project is underestimated, which makes the real value of the venture capital project can not be effectively revealed, the venture capital is difficult to smooth out, the market efficiency is low; because of the venture capital problem, the repurchase is also difficult; liquidation is only the means to choose when the venture enterprise is bankrupt. It is difficult to realize the maximization of capital value of venture investors. The core of the problem is to reveal the real value of venture enterprises or projects under asymmetric information, and to make the capital realize the maximum value of the venture capital, the venture capital is successfully withdrawn and the capital appreciation of the venture capital is achieved. The purpose of recycling.
Because auction is essentially an effective means and value disclosure mechanism to allocate resources under asymmetric information, auction theory can solve the problem of maximizing the value of goods and optimizing the allocation of resources, and then evade the moral hazard and adverse selection caused by asymmetric information when venture capital exit. In the view of the square, two kinds of mathematical model of the multi attribute auction mechanism of the venture capital equity withdrawal - the unified price auction model and the discriminatory price auction model are established to maximize the equity value of the venture capital. In this paper, we use the multi utility theory and use the multi utility theory to select the bidder in the auction system in two stages. In the first stage, the multi utility theory is used for comprehensive scoring to eliminate some low quality bidder. In the second stage, a single attribute item is auctioned on the most important price attribute, and the risk is put in accordance with the risk. The most important attribute of the equity withdrawal from the auction, the price attribute, divides the auction into a unified price auction and a discriminatory price auction, on the basis of which a corresponding mathematical model is established.
In view of the characteristics of the equity auction of venture capital, such as the intelligent behavior of the bidder, the mutual learning, the information transmission, and the bidding strategy of the bidder, the multi-agent system introduces the auction mechanism to solve the problem of the maximization of the value of the equity auction and the allocation of resources. It is assumed that the main body of the intelligence (Agent) is to buy. The bidder, in the open, dynamic multi-agent system, the self-interest body has its own goals and preferences, through observation, negotiation and reasoning, to determine what kind of bidding behavior strategy. The single subject can not accomplish the goal independently, and need the participation and help of other subjects, that is to say, we need to jointly solve the advantages through multiple main forces. According to the theory of auction theory, when the bidder adopts the strategy of telling the truth, it is the best strategy for the bidder, and the auctioneer is also expected to get the maximum benefit, and the auction system will reach a balance. But in the case of the venture capital stock withdrawal from the associated value auction, the competitor with limited rationality should be excluded. In the case of noise interference, it is difficult to say the truth, so using the learning optimization mechanism of intelligent algorithm can make the bidder achieve the optimal solution with the optimization of the intelligent group, the auction system reaches the equilibrium, which makes the auctioneer with the venture capital stock exit expect the maximum benefit. This paper uses the particle swarm optimization (PSO) to optimize the venture capital stock. The bidding strategy of the bidder is unified by the mathematical model of the price auction, and the genetic algorithm (GA) is used to optimize the bidding strategy of the bidders in the discriminatory price auction model when the risk investment is withdrawn.
The model based on the associated value auction system and the bidding strategy of the bidder can be considered as a dynamic and complex system. The traditional mathematical methods and experimental methods are difficult to solve. In this paper, the method of complex system is used to solve the optimization of the quotation of complex auction system and the maximization of the expectation of the auctioneer. Multi subject modeling and simulation is a very flexible modeling and simulation technology. The essential feature is to establish the conceptual model of the system by multi-agent perspective. In this paper, it is based on the multi-agent perspective to describe the exit auction of risk investment. In the auction bidding process, the bidder is the auction bidding based on the associated value, due to the pre bidding owners. The thought and behavior of the bidders in reality are complex. They have intelligence, adaptability, learning and interactivity. There must be a noise form of information communication channel between the subjects. The subject will get some other people's bidding information so as to maximize their own interests. This complex selection process meets the characteristics of many subjects. Therefore, in this paper, the multi-agent simulation platform Netlogo is used to verify the bidding strategy of each subject, the unified price auction model and the discriminatory price auction model. The main flow of the auction system is the actual system, the conceptual model, the simulation model, the simulation result and the conclusion. Finally, the conclusion is used to verify and verify the results. It is correct and reasonable to modify the venture capital withdrawal auction model.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：TP18;F832.51;F724.59

【相似文献】