考虑背景风险与社会责任的投资组合模型及算法研究

发布时间：2018-01-17 02:26

本文关键词：考虑背景风险与社会责任的投资组合模型及算法研究　出处：《华南理工大学》2016年博士论文　论文类型：学位论文

【摘要】：近年来,由于金融市场风险的多样化和投资活动的复杂化,投资者的风险承受能力和财务收支状况日益受到关注。由于社会的不断进步,经济水平的快速发展,企业社会责任(简称:社会责任)的缺失或滞后与社会、环境、伦理的矛盾不断加剧。在Markowitz提出M-V模型之后,很多学者根据他的思想在此基础上不断扩展。目前,很多学者的研究集中在证券的财务收益率均值、方差等指标上,而对于其他内容的研究则相对较少,例如投资者面临的背景资产风险(简称为:背景风险)和上市公司的社会责任等非财务指标也是影响投资活动收益与风险的重要因素。本文运用概率论、模糊数学、优化理论以及智能算法研究考虑背景风险和社会责任的投资组合问题,构造和求解对应的投资组合模型。本文的主要研究工作以及创新点概括如下:1、分别构造了一个考虑背景风险和弹性增量的随机投资组合模型和一个考虑背景风险和弹性增量的模糊投资组合模型,设计了一个改进的混沌果蝇优化算法。传统的投资组合模型没有考虑背景风险,而且研究的对象往往是投资组合收益、方差等指标的存量及其相关问题,很少研究投资组合中这些指标的增量问题及其相关问题。因此,本文分别建立了一个考虑背景风险和弹性增量的投资组合模型,涉及投资组合收益、方差、和流动性三个指标的存量和增量,然后,定义了这三个指标的隶属函数,研究了背景风险、三个指标的存量及其增量对投资者满意度的影响。最后,引入混沌算子初始化果蝇种群,提出了一个改进的混沌果蝇优化算法。2、分别构造了一个考虑社会责任的期望值多层次投资组合模型和一个考虑社会责任的相关机会多层次投资组合模型,设计了一个基于云模型的混合果蝇优化算法。现实的投资决策,一般来说,是一个分层的决策过程:决策者往往不是一个,而是同时有多个决策者,包括一个上级和若干个下级。一般而言,上级与他(她)的下级都有独立的决策变量和决策目标。上级依据其行为制约下级,而下级下属则有充分权限如何行动实现决策目标,这些行为又将制约他(她)的上级和下级。因此,本文构造一个考虑社会责任的期望值多层次投资组合模型和一个考虑社会责任的相关机会多层次投资组合模型。最后,利用遗传算法和神经网络求解上级和下级之间的纳什均衡。3、分别构造了一个考虑企业社会责任和风险厌恶系数的多阶段投资组合模型和一个考虑企业社会责任和基数约束的投资组合模型,设计了一个云模型混合果蝇算法。在实际的投资活动中,对于投资者而言,需要在每一个阶段随着市场环境的变化调整投资组合头寸,到最后一个投资阶段结束时,实现总收益最大化或者总风险的最小化,考虑到证券交易过程中产生的费用,例如印花税和手续费,投资者不可能同时持有很多支证券。因此,本文分别构造了一个考虑社会责任和风险厌恶系数的多阶段投资组合模型和一个考虑社会责任和基数约束的投资组合模型。最后,借鉴云算子、交叉算子和变异算子,构造了一个云模型混合果蝇算法。4、构造了一个考虑社会责任和背景风险的随机多目标投资组合模型和一个考虑社会责任和背景风险的模糊多目标投资组合模型。由于投资风险来源的多元化及社会、环境、伦理等问题的复杂化,投资者需要同时考虑自身的背景风险和证券发行人-企业的社会责任。其次,金融市场中不只有低阶矩风险,也有偏度、峰度等高阶矩风险。因此,本文分别构造了一个考虑社会责任和背景风险的随机多目标投资组合模型和一个考虑社会责任和背景风险的模糊多目标投资组合模型。另外,在概率测度下和模糊测度下,分别推导出了考虑背景风险的投资组合偏度公式和峰度公式。最后,本文借鉴非占优排序算子、交叉算子和变异算子,提出了一个非占优排序的混合果蝇多目标优化算法。
[Abstract]:In recent years, due to the complexity and diversification of investment activities and financial market risk, investors risk tolerance and financial situation is getting more and more attention. Due to the progress of the society, the rapid development of economic level, corporate social responsibility (referred to as: social responsibility) in the absence or lag with society, environment, ethics contradiction intensified after the M-V model proposed in Markowitz, many scholars on the basis of his thoughts on the basis of continuous expansion. At present, many scholars have focused on the financial income rate, variance index, and the study of other content is relatively small, such as the background of asset risk faced by investors (hereinafter referred to as the important background: risk) factors of social responsibility and other non financial indicators of listed companies and also affect the benefits and risks of investment activities. Based on probability theory, fuzzy mathematics, optimization theory and Research on intelligent algorithm considering the problem of portfolio risk background and social responsibility, portfolio model construction and the corresponding solution. The main research work and innovations are as follows: 1. A background of considering risk and elastic incremental stochastic portfolio model and a fuzzy model of portfolio investment risk and the increment of elastic background are constructed, designed an improved chaos optimization algorithm. The Drosophila traditional portfolio model without considering the background of risk, and the research objects are often portfolio returns, inventory variance index and its related problems, and problems related to incremental problem is rarely studied these indicators in the portfolio. Therefore, this paper establishes a portfolio model of background risk and elastic increment, involving portfolio returns, variance, and liquidity three indicators and stock Then, incremental, the membership function of the three indicators are defined, research background of risk, stock and increment of three indicators of investor satisfaction. Finally, introducing chaos operator initialization of Drosophila melanogaster population, the paper proposes an improved chaos optimization algorithm.2 Drosophila respectively, construct a consideration of social responsibility expectation the value of multi-level portfolio model and a consideration of the social responsibility of the relevant opportunity multi-level portfolio model, designed a hybrid Drosophila optimization algorithm based on cloud model. The reality of the investment decision, in general, is a hierarchical decision making process: decision makers tend to not one, but at the same time there are many decision makers, comprising an upper and a number of subordinate. In general, he (she) superior and subordinates have independent decision-making variables and goals. On the basis of the superior and subordinate behavior constraints, under the full level How to realize the goal of decision rights action, these actions will restrict his (her) superiors and subordinates. Therefore, this paper constructs a social responsibility expectation of multi-level portfolio model and a consideration of the social responsibility of the relevant opportunity multi-level portfolio model. Finally, using the Nash equilibrium between genetic algorithm and neural.3 the network is superior and subordinate, a multi period portfolio model considering corporate social responsibility and the risk aversion coefficient and a consideration of corporate social responsibility and the cardinality constrained portfolio model, designed a hybrid cloud model Drosophila algorithm. In the actual investment activities, for investors, need at every stage along with the change of market environment to adjust portfolio positions, to the end of the last stage of investment, to achieve a minimum of maximum total revenue or total risk, To account in securities trading costs, such as taxes and fees, investors can not hold a lot of funds at the same time. Therefore, this paper constructs a multi period portfolio model considering the social responsibility and the risk aversion coefficient and a consideration of the social responsibility constraint and cardinality term portfolio model. Finally, from the cloud operator, crossover operator and mutation operator, we construct a hybrid cloud model of Drosophila.4 algorithm, constructed a stochastic multi objective portfolio model considering the social responsibility and the background risk and a fuzzy multi-objective portfolio model of social responsibility and background risk. Because of the diversified sources of investment risk and society. The environment, complex ethical issues, investors need to consider its own background and the risk of the securities issuer - corporate social responsibility. Secondly, not only in the financial market Low order moments have risk, skewness, kurtosis and higher moments risk. Therefore, this paper constructs a stochastic multi-objective portfolio model considering the social responsibility and the background risk and a fuzzy multi-objective portfolio model of social responsibility and background risk. In addition, the probability measure and fuzzy measure respectively, and deduces the background risk portfolio skewness and kurtosis formula formula. Finally, based on non dominated sorting operator, crossover operator and mutation operator. The proposed multi-objective optimization algorithm for a non dominated sorting of mixed fruit flies.

【学位授予单位】：华南理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F830.59;F224

【相似文献】