Service contracts: A stochastic model

There is a growing trend to outsource maintenance where equipment failures are rectified by an external agent under a service contract. The agent's profit is influenced by many factors—the terms of the contract, equipment reliability, and the number of customers being serviced. The paper develops a stochastic model to study the impact of these on the agent's expected profit and the agent's optimal strategies using a game theoretic formulation.     

Self‐organized trail systems in groups of humans

We have developed an experimental platform for studying the trail systems that spontaneously emerge when people are motivated to take advantage of the trails left by others. In this virtual environment, the participants' task is to reach randomly selected destinations while minimizing travel costs. The travel cost of every patch in the environment is inversely related to the number of times the patch was visited by others. The resulting trail systems are a compromise between people going to their destinations and going where many people have previously traveled. We compare the results from our group experiments to the Active Walker model of pedestrian motion from biophysics. The Active Walker model accounted for deviations of trails from the beeline paths, the gradual merging of trails over time, and the influences of scale and configuration of destinations on trail systems, as well as correctly predicting the approximate spatial distribution of people's steps. Two deviations of the model from empirically obtained results were corrected by (1) incorporating a distance metric sensitive to canonical horizontal and vertical axes, and (2) increasing the influence of a trail's travel cost on an agent's route as the agent approaches its destination. © 2006 Wiley Periodicals, Inc. Complexity 11: 43–50, 2006     

View-based explicit knowledge

In the information-based definition of knowledge, an agent is said to know α at a state s if α is true in all states that look the same as s to that agent. However, in systems where an agent's view of the system is partial, even if a state s′ may be logically indistinguishable from a state s, s′ may not be visible from s. For instance, in a distributed system, all global states in which the agent's local state does not change look the same to that agent but this set of global states may not be accessible because the agent may not even be aware of the existence of many agents in the network. We propose a logic of explicit knowledge built on agents' views and show it to be decidable.     

Mover‐stayer model with covariate effects on stayer's probability and mover's transitions

A discrete time Markov chain assumes that the population is homogeneous, each individual in the population evolves according to the same transition matrix. In contrast, a discrete mover‐stayer (MS) model postulates a simple form of population heterogeneity; in each initial state, there is a proportion of individuals who never leave this state (stayers) and the complementary proportion of individuals who evolve according to a Markov chain (movers). The MS model was extended by specifying the stayer's probability to be a logistic function of an individual's covariates but leaving the same transition matrix for all movers. We further extend the MS model by allowing each mover to have her/his covariates dependent transition matrix. The model for a mover's transition matrix is related to the extant Markov chains mixture model with mixing on the speed of movement of Markov chains. The proposed model is estimated using the expectation‐maximization algorithm and illustrated with a large data set on car loans and the simulation.     

A Multi-Layered Potential Field Method for Multi-Agent Navigation,Searching and Tracking

Collaborating multi-agent systems can handle complex tasks with several or changing mission objectives. We developed a potential field method that allows various information layers to influence the control over a group of vehicles. The gradient of the potential field is the driving force for local action, whereas the global waypoint is determined by the minimum of the agent's potential field. The driving force to the global waypoint is a virtual spring-mass-damper system that pulls the agent towards its waypoint, restricted by the local gradient of the agent's potential field. (© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Dominant strategy mechanisms for contract auctions with risk aversion and moral hazard

Within the class dominant strategy incentive compatible mechanisms, we show that there exists an optimal contracting mechanism for the principal for a version of the incomplete information principal-agent problem in which several agents compete for a contract and the principal selects an agent via a contract auction. In our auction model, we assume that the principal and the agents are risk averse, and we allow for uncountably many agent types. We also assume that the principal's probability measure over type profiles in such that correlation between agent's types is possible. Thus, we do not require that agents' types be independently distributed. Finally, we impose limited liability constraints upon the set of contracts. Due to the nature of the individual rationality and incentive compatibility constraints, the existence problem is nonstandard and novel existence arguments are required. We prove existence using a measurable selection result and a new notion of compactness called K-compactness.     

代理人在职消费对动态多任务激励契约的影响

将代理人的在职消费行为引入到动态多任务委托代理框架中,构造了代理人在职消费行为下的两阶段多任务模型,分析了代理人在职消费行为对动态多任务激励契约的影响.研究结果表明:一是任务为两阶段时,无论代理人有无在职消费行为,代理人的努力程度随着时间均呈上升趋势,这就表明当委托人在设计契约时,如果委托人期望代理人在第一阶段的努力水平不低于第二阶段的努力水平,就需要适当提高第一阶段的业绩薪酬系数;二是代理人在职消费自利行为并不一定会提高自身的努力程度,需要依据在职消费行为对绩效的影响情形来具体分析;三是在两阶段内,代理人存在在职消费时,委托人可适当降低业绩薪酬系数.     

在线知识社区中基于个体异质性的知识与社会系统协同演化研究

在Web 2.0环境下,在线知识社区(Online Knowledge Community, OKC)已成为人们进行学习、利用、分享、传播知识的重要平台。然而,现有关于OKC的研究较多关注其中社会系统对个体知识建构、群体知识建构的影响,缺少从OKC演化的视角,以用户为基础对OKC中社会系统与知识系统间双向交互影响规律的研究。为此,本文将个体异质性作为影响因素引入到OKC的研究中,在一个开放的在线知识社区中模拟个体的观点、认知结构以及个体间形成的社会关系与知识系统的共同演化。将个体间以信任为基础形成的关系网络视为OKC中的社会系统,将群体依据自身认知结构表达出的全部观点视为社区中的知识系统,进而采取了社会科学计算实验的研究方法构建了一个OKC中社会系统与知识系统协同演化模型,并在模型中引入了个体的内在和外在异质性作为影响因素。研究结果表明OKC中个体的学习能力、容忍度、活跃度等个体因素,对其中知识观点的传播与社区中社会网络的演化影响显著。本研究希望能够为OKC中知识的自组织、序化的研究提供理论支持。     

The elastic generalized assignment problem

The generalized assignment problem (GAP) has been studied by numerous researchers over the past 30 years or so. Simply stated, one must find a minimum-cost assignment of tasks to agents such that each task is assigned to exactly one agent and such that each agent's resource capacity is honoured. The problem is known to be NP-hard. In this paper, we study the elastic generalized assignment problem (EGAP). The elastic version of GAP allows agent resource capacity to be violated at additional cost. Another version allows undertime costs to be assessed as well if an agent's resource capacity is not used to its full extent. The EGAP is also NP-hard. We describe a special-purpose branch-and-bound algorithm that utilizes linear programming cuts, feasible solution generators, Lagrangean relaxation and subgradient optimization. We present computational results on a large collection of randomly generated ‘hard’ problems with up to 4000 binary variables.     

Space-time dynamic of normalized doxatons: automata models of pathological collective mentality

The paper aims to investigate influences of norms on a space-time dynamic of stirred (solutions) and non-stirred (lattices) collectives of very simple believing agents. A mental state of an agent is characterized by agent's belief in some proposition and a truth value of the proposition in agent's local vicinity. Every agent of the collective updates its belief depending on its current belief, belief of its closest neighbours and truth value of the proposition in neighbourhood of the neighbours. Each agent is represented by a finite automaton – a so-called doxaton [A. Adamtzky, Appl. Math. Comput., in press]. The doxaton takes five doxastic states, derived from the belief: knowledge, doubt, misbelief, delusion and ignorance. In the above mentioned reference, we defined a binary composition of doxastic states and investigated its algebraic structure. The composition itself is not deterministic; however, it can be made deterministic by applying norms. The norms are expressed in a priority order on the doxastic states. Space-time evolution of the solutions and the lattices of doxatons is studied in computer experiments to understand influences of norms and initial conditions on the behaviour of abstract collectives of simple agents. Diffusion and reaction of doxastic states are explored as well as formation of stationary patterns of the doxastic states.     

Robust learning in social networks via matrix scaling

The influence vanishing property in social networks states that the influence of the most influential agent vanishes as society grows. Removing this assumption causes a failure of learning of boundedly rational dynamics. We suggest a boundedly rational methodology that leads to learning in almost all networks. The methodology adjusts the agent's weights based on the Sinkhorn-Knopp matrix scaling algorithm. It is a simple, local, Markovian, and time-independent methodology that can be applied to multiple settings.     

改进GA机制融合的二进制蜂群算法研究

为了克服人工蜂群算法蜜源更新过程中的随机性并保留蜜源中个体序列合理的组合形式,通过分析基本蜂群算法更新公式的机理,提出一种改进GA(Genetic A1gorithm)机制融合的二进制蜂群算法.算法以二进制编码,首先依概率对任意两蜜源进行"去同存异"操作后随机排列,将排列结果放入到其中某个体中形成新个体.然后依概率进行二进制个体的"翻转"操作,上述两种操作从其本质上相当于GA的类交叉和类变异操作;其次利用GA机制收敛性的证明方式在理论上证明算法是收敛的.最后通过应用不同特性的多维基准函数和算法之间的比较验证改进蜂群算法具有良好的收敛能力和鲁棒性.     

Learning management knowledge for manufacturing systems in the early stages using time series data

Acquiring knowledge in manufacturing systems in the early stages always has a challenging task due to the lack of sufficient data. This makes it hard for the derived management model to reach a reliable and stable level. Li and Lin (2006) developed a useful method that can deal with the problem of knowledge acquisition based on a small data set. However, this method assumes all data are collected at the same time, since they treat the data set as a source (from one population) of a priori knowledge for learning. In fact, instead of being a random data set, these collected data can be time dependent, that is, they tend to be a sequence of observations, occurring at different times. The consideration of this dependence property in the data will benefit the knowledge acquisition in the early stages by expanding the learning model from an independent model to a dependent model. This research expanded the intervalized kernel density estimator (IKDE) presented in Li and Lin (2006) to a more general form to improve the learning model in the early stages. The general model, called GIKDE, joints the concepts of time series and stochastic processes in order to deal with both independent and dependent data sets. The Virtual Sample Generation process based on GIKDE was also developed to produce extra information for expediting the learning. Results obtained from the application of the model to a control charts data, using a back-propagation neural network as the learning tool, show that this unique approach is an effective method of knowledge acquisition for a manufacturing system in the early stages.     

Common due date scheduling with autonomous and induced learning

Up to now the few existing models, that consider learning effects in scheduling, concentrate on learning-by-doing (autonomous learning). But recent contributions to the literature on learning in manufacturing organizations emphasize the important impact of proactive investments in technological knowledge on the learning rate (induced learning). In the present paper, we focus on a scheduling problem where the processing times decrease according to a learning rate, which can be influenced by an initial cost-inducing investment. Thus we have integrated into our model both aspects of learning––autonomous and induced––thereby highlighting the management's responsibility to invest in technological knowledge enhancement. We have been able to derive some structural properties of the problem and present a polynomially bound solution procedure which optimally solves the problem by using these properties. The optimal solution to the scheduling problem contains––of course–– information on the optimal level of proactive investments in learning.     

Integration of the grey relational analysis with genetic algorithm for software effort estimation

Accurate estimates of efforts in software development are necessary in project management practices. Project managers or domain experts usually conduct software effort estimation using their experience; hence, subjective or implicit estimates occur frequently. As most software projects have incomplete information and uncertain relations between effort drivers and the required development effort, the grey relational analysis (GRA) method has been applied in building a formal software effort estimation model for this study. The GRA in the grey system theory is a problem-solving method that is used when dealing with similarity measures of complex relations. This paper examines the potentials of the software effort estimation model by integrating a genetic algorithm (GA) to the GRA. The GA method is adopted to find the best fit of weights for each software effort driver in the similarity measures. Experimental results show that the software effort estimation using an integration of the GRA with GA method presents more precise estimates over the results using the case-based reasoning (CBR), classification and regression trees (CART), and artificial neural networks (ANN) methods.     

Preparation for Practice: Elementary Preservice Teachers Learning and Using Scientific Classroom Discourse Community Instructional Strategies

Despite historical national efforts to improve elementary science education, science instruction continues to be marginalized, varying by state. This study was designed to address the ongoing challenge of educating elementary preservice teachers (PSTs) to teach science. Elementary PSTs are one of the science education community's major links to schools and science education reform. However, they often lack a strong background in science, knowledge of effective science teaching strategies, and consequently have low confidence and self‐efficacy. This investigation explored the initial learning of elementary PSTs using an interdisciplinary model of a scientific classroom discourse community during a science methods course. Findings post‐methods course suggested that the PSTs gained confidence in how to teach inquiry‐based elementary science and recognized inquiry‐based science as an effective means for engaging student learning. Additionally, PSTs embraced the interdisciplinary model as one that benefits students' learning and effectively uses limited time in a school day.     

Emergent learning behaviour in a simulated organization faced with tasks requiring team effort

We conceptualize organizational learning as a result of the collective learning behaviour of knowledge agents in an organization. Each agent provides a range of attributes that may be required to perform organizational tasks. We devised a computational model consisting of three processes to simulate an organization's response to performing repeated tasks: (1) Expert Selection Process for selecting the winner knowledge agent or lead agent; (2) Plan Formation Process for deciding what additional attributes are needed, but not possessed by the winner expert agent, and iteratively selecting further agents with the needed attributes until the task can be accomplished by the combined attributes of the ‘coalition of agents’ so formed; and (3) Capital Modification Process for rewarding participating agents according to the success of their combined organizational performance. We observed the simulated results for different combinations of three levels of task difficulty (requiring, respectively, 5, 10 and, 15 different attributes, each at a sufficient level in the coalition or team to complete the task), and three levels of selection, during plan formation, for knowledge agent performance (the extent to which selection favours knowledge agents with much capital or large strength versus knowledge agents without much capital or large strength). The simulated organization exhibited aspects of both single loop and double loop learning, in repeatedly performing the same task, and ‘learning to perform the task’ with the smallest possible team.     

Security design and firm dynamics under long‐term moral hazard

This paper addresses a moral hazard problem in which the agent's actions affect the future profits of the firm. The optimal contract can be implemented through the issuance of variable coupon debt and purchase of fixed‐coupon debt. Consequently, the resulting capital structure acts as a hedge for the firm, reducing underinvestment costs in bad states of nature and controlling overinvestment incentives in good ones. However, owing to asymmetric information between the firm's manager and investors, this hedge is only partial. The firm's investments vary with cash flows, disclosing the agent's asymmetric information to the principal. Copyright © 2016 John Wiley & Sons, Ltd.     

Autonomous vehicle navigation using evolutionary reinforcement learning

Reinforcement learning schemes perform direct on-line search in control space. This makes them appropriate for modifying control rules to obtain improvements in the performance of a system. The effectiveness of a reinforcement learning strategy is studied here through the training of a learning classifier system (LCS) that controls the movement of an autonomous vehicle in simulated paths including left and right turns. The LCS comprises a set of condition-action rules (classifiers) that compete to control the system and evolve by means of a genetic algorithm (GA). Evolution and operation of classifiers depend upon an appropriate credit assignment mechanism based on reinforcement learning. Different design options and the role of various parameters have been investigated experimentally. The performance of vehicle movement under the proposed evolutionary approach is superior compared with that of other (neural) approaches based on reinforcement learning that have been applied previously to the same benchmark problem.