Utility Function Programs and Optimization over the Efficient Set in Multiple-Objective Decision Making

Natural basic concepts in multiple-objective optimization lead to difficult multiextremal global optimization problems. Examples include detection of efficient points when nonconvexities occur, and optimization of a linear function over the efficient set in the convex (even linear) case. Assuming that a utility function exists allows one to replace in general the multiple-objective program by a single, nonconvex optimization problem, which amounts to a minimization over the efficient set when the utility function is increasing. A new algorithm is discussed for this utility function program which, under natural mild conditions, converges to an -approximate global solution in a finite number of iterations. Applications include linear, convex, indefinite quadratic, Lipschitz, and d.c. objectives and constraints.     

A distributed Newton algorithm for network utility maximization in wireless ad hoc networks

The existing distributed Newton algorithm for network utility maximization cannot be directly applied into the wireless ad hoc networks, as it does not consider the wireless link capacity variation and transmission power consumption. Regarding these, a new joint network utility maximization problem is formulated to optimize the wireless ad hoc network resource utility. The distributed Newton algorithm is implemented to solve for the dual variables and Newton directions by using the local information at each session source and link, respectively. Furthermore, a new iterative method is proposed to improve the convergence rate of the existing matrix‐splitting method. Simulation results validate the efficiency and efficacy of the proposed distributed Newton algorithm for wireless ad hoc network utility maximization.     

基于多天线波束赋形的CRN分布式上行功率控制算法

针对工作于underlay模式的认知无线网络（CRN,Cognitive Radio Network）上行功率控制问题,本文提出一种基于多天线波束赋形,由认知基站和认知用户联合优化的分布式上行功率控制算法.联合优化的具体步骤为认知基站通过求解最大广义特征值问题完成多天线波束赋形优化;认知用户先将非线性功率优化问题转换为几何规划凸优化问题,再使用梯度法完成分布式发送功率优化;认知基站和认知用户交替优化,实现网络效用最大化.数值仿真显示,同只优化认知用户功率的上行功率控制算法相比,认知基站和认知用户联合优化的上行功率控制算法不仅能得到更大的网络效用值,而且对主用户的干扰具有鲁棒性.     

A new learning automata based approach for increasing utility of service providers

Utility is an important factor for serviceproviders, and they try to increase their utilities through adopting different policies and strategies. Because of unpredictable failures in systems, there are many scenarios in which the failures may cause random losses for service providers. Loss sharing can decrease negative effects of unexpected random losses. Because of capabilities of learning automata in random and stochastic environments, in this paper, a new learning automaton based method is presented for loss sharing purpose. It is illustrated that the loss sharing can be useful for service providers and helps them to decrease negative effect of the random losses. The presented method can be used especially in collaborative environments such as federated clouds. Results of the conducted experiments show the usefulness of the presented approach to improve utility of service providers.     

Differentiated service delivery in cooperative IP-based broadcast and mobile telecommunications networks

Email: angelo.centonza{at}brunel.ac.uk Corresponding author. Email: thomas.owens{at}brunel.ac.uk Email: john.cosmas{at}brunel.ac.uk^¶ Email: y.h.song{at}brunel.ac.uk Received on 26 November 2005. Accepted on 20 March 2007. The evolution of wireless systems has led recently to the deploymentof cooperative network infrastructures where networks basedon different technologies cooperate together to offer innovativeservices which the networks individually could not offer. Cooperativenetwork infrastructures are hybrid systems. Examples of hybridsystems are already in use in areas as diverse power systemsand locomotion systems. It is very important to apply efficientsystem management techniques to hybrid systems that produceadvantageous business case scenarios for each participatingsystem and efficient use of the available resources. In thispaper, scenarios are investigated where cooperation betweenan integrated project (IP)-based broadcast network and an IP-basedmobile telecommunications network has the potential to allowservices to be provided to mobile users which would not be economicalto offer over either an IP-based broadcast network or an IP-basedmobile telecommunications network alone. The paper presentsa novel technique for determining which network to use to deliversuch services at a given point in time. The application of thistechnique in appropriate scenarios has the potential to generateadditional income for both the IP-based broadcast network operatorand the IP-based mobile telecommunications network operatorin an infrastructure where the operators cooperate to offerinnovative services. The paper explains the construction ofutility functions for cooperative IP-based broadcast and mobiletelecommunications networks. These utility functions are thenused to provide results that enable the efficiency of the managementof the network to be assessed in terms of the utility volumegenerated by the innovative services provided.     

劳动者行为分析

以边际分析和效用理论研究人的劳动行为,发现人的行为规律性及动力因素与社会偏好的联系,试图对人的一般社会行为研究提供一种定量与定性结合的方法     

变折现率下带含糊厌恶与预期的最优投资研究

本文采用折现率为时间的函数下的递推多先验效用,研究Merton模型在带预期条件下的最优消费和投资组合决策问题,其中含糊与风险是有区别的.在幂效用函数情形下,刻画了投资者最优投资决策,表明了含糊厌恶和预期对最优投资的影响.最优投资组合决策由倒向随机微分方程和Malliavin导数导出.     

基于成本效用分析的电能替代实证研究

近年来的雾霾让人们逐渐意识到环境保护的重要性,国网公司基于我国“多煤、少气、贫油”的国情,提出了以电代煤的电能替代方案。本文以在电力和煤炭的使用过程中可获得的热值作为效用,以使用过程中所需要的年费用作为成本,建立了电能替代的成本效用模型,计算出实现电力和煤炭相互替代的排污费临界值,并通过一个算例进行了实证分析。分析结果表明:电价、单位电力排污量、燃煤设备寿命对排污费临界值有正向影响;煤价、单位燃煤排污量、电力设备寿命对其有负向影响。最后,在分析的基础上给出了电能替代的政策建议。     

Alternative formulations of a combined trip generation,trip distribution,modal split,and trip assignment model

The traditional four-step model has been widely used in travel demand forecasting by considering trip generation, trip distribution, modal split and traffic assignment sequentially in a fixed order. However, this sequential approach suffers from the inconsistency among the level-of-service and flow values in each step of the procedure. In the last two decades, this problem has been addressed by many researchers who have sought to develop combined (or integrated) models that can consider travelers’ choice on different stages simultaneously and give consistent results. In this paper, alternative formulations, including mathematical programming (MP) formulation and variational inequality (VI) formulations, are provided for a combined travel demand model that integrates trip generation, trip distribution, modal split, and traffic assignment using the random utility theory framework. Thus, the proposed alternative formulations not only allow a systematic and consistent treatment of travel choice over different dimensions but also have behavioral richness. Qualitative properties of the formulations are also given to ensure the existence and uniqueness of the solution. Particularly, the model is analyzed for a special but useful case where the probabilistic travel choices are assumed to be a hierarchical logit model. Furthermore, a self-adaptive Goldstein–Levitin–Polyak (GLP) projection algorithm is adopted for solving this special case.     

Optimal investment and reinsurance of an insurer with model uncertainty

We introduce a novel approach to optimal investment–reinsurance problems of an insurance company facing model uncertainty via a game theoretic approach. The insurance company invests in a capital market index whose dynamics follow a geometric Brownian motion. The risk process of the company is governed by either a compound Poisson process or its diffusion approximation. The company can also transfer a certain proportion of the insurance risk to a reinsurance company by purchasing reinsurance. The optimal investment–reinsurance problems with model uncertainty are formulated as two-player, zero-sum, stochastic differential games between the insurance company and the market. We provide verification theorems for the Hamilton–Jacobi–Bellman–Isaacs (HJBI) solutions to the optimal investment–reinsurance problems and derive closed-form solutions to the problems.