THE LEARNING CURVE: HISTORICAL REVIEW AND COMPREHENSIVE SURVEY

The use of the learning curve has been receiving increasing attention in recent years. Much of this increase has been due to learning curve applications other than in the traditional learning curve areas. A comprehensive survey of developments in the learning curve area has never been published. The closest thing to a survey was by Asher in 1956. His study focused exclusively on military applications during and immediately after World War II. This paper summarizes the learning curve literature from World War II to the present, emphasizing developments since the study by Asher. Particular emphasis is given to identifying the new directions into which the learning curve has made recent inroads and identifying fruitful areas for future research.     

Adding life cycles to learning curves

The price-oriented experience curve has been extensively utilized as a model in corporate planning. However, managers exercise a greater degree of control over cost than they do over price. The use of the cost-oriented learning curve together with the product life cycle model in strategic planning is explored in this paper. The Ford Motor Company's experience with the Model-T, which was the first mass-produced automobile in the U.S., is used to reinforce the fact that the simultaneous use of the product life cycle and the learning curve would have provided Ford management with better planning information. With this information, Ford could have avoided the subsequent disastrous results which it experienced.     

A COMPARISON OF LEARNING CURVE ANALYSIS AND MOVING AVERAGE RATIO ANALYSIS FOR DETAILED OPERATIONAL PLANNING*

This study tests the use of learning curve analysis for production planning at the detailed component level under various conditions, represented by factors of product turnover rate, learning rate, variance levels, and planning horizon length. It also presents an alternative to learning curve analysis that considers aggregation of cost data across time. This alternative is periodic revision of standard cost data using moving average forecasts to reflect productivity trends. Results of this study indicate that in most circumstances a moving average analysis can provide better estimates of short-term, detailed component operations costs than either a learning curve analysis or a standard analysis.     

Learning to learn and productivity growth: Evidence from a new car-assembly plant

This paper models learning by experience beyond the experience curve, including the possibility of “learning to learn”: the pace of learning increases over time by building on what has already been learned. We compare the extended deterministic learning model with Jovanovic and Nyarkos' [26] stochastic learning. The theoretical models are tested with data on the total factor productivity of a car-assembly plant in its first months of operation. We find that the deterministic “mixed learning model”, where the speed of learning is equal to a constant plus a learning to learn effect, is the one that best fits the empirical data. The mixed learning model results in a time pattern of total factor productivity growth, first increasing and later decreasing, different from the always decreasing rate of growth of the learning curve, opening new perspectives on the study of learning by experience.     

Reducing Customer Dissatisfaction: How Important is Learning to Reduce Service Failure?

As service failures are inevitable, firms must be prepared to recover and learn from service failures. Yet, the majority of customers are still dissatisfied with the way firms resolve their complaints. Can learning to reduce service failures reduce customer dissatisfaction, and to what extent are such reductions sustainable? Previous research showed that organizational learning curves for customer dissatisfaction (i) follow a U‐shaped function of operating experience and (ii) are heterogeneous across firms. In this paper, I tease out where the U‐shaped learning‐curve effect and learning‐curve heterogeneity originate: service failure or customers' propensity to complain with a third party given the occurrence of a service failure. Using quarterly data for nine major US airlines over 11 years, I find that the U‐shaped learning‐curve effect and the learning‐curve heterogeneity originate in the propensity to complain. In the long term, reductions in service failure did not translate in sustainable reductions in customer dissatisfaction. Customers' propensity to complain eventually went up. Managing the propensity to complain provides more opportunity for a firm to distinguish itself from competitors.     

AN ECONOMIC PRODUCTION QUANTITY MODEL WITH LEARNING AND FORGETTING CONSIDERATIONS

The focus of this work is on the effects of learning on economic production quantity in batch production systems. We assumed that both unit variable manufacturing time and setup time follow a learning curve. We modified the classical Economic Production Quantity model to incorporate these two types of learning phenomena. We also incorporated the forgetting effect in our model so that a fraction of the learning is lost between consecutive lots. We developed a dynamic program to obtain the optimal solution to the problem. We investigated the nonincreasing lot size property and used it to improve the efficiency of our dynamic program. We consider a special case of the model in which all lot sizes are assumed equal. After theoretical treatment, we carried out a computational study of the effect of assuming equal lot sizes on the optimal solutions. The results of our examples strongly indicate that the assumption of equal lot sizes not only simplifies the determination of the optimal solutions, but also provides close approximations to the optimal solutions.     

MODELING LEARNING CURVE AND LEARNING COMPLEMENTARITY FOR RESOURCE ALLOCATION AND PRODUCTION SCHEDULING*

Research on learning effects in mathematical programming models for optimum resource allocation has called attention to the difficulty in solving such models in their original nonlinear form. In this paper, systematically varying sizes of linear segments are designed to approximate productivity changes along the learning curve, and a single separable linear programming model is developed. With production complementarity and learning transmission between products, a more realistic resource allocation and production scheduling problem emerges. Two cases of learning transmissions are considered, and the model design process, which defines a decision problem that can be solved by a simplex algorithm, is demonstrated.     

Learning and Bayesian updating in long cycle made-to-order (MTO) production

We model production planning for made-to-order (MTO) manufacturing by choosing production rate to minimize expected discounted cost incurred up to a promised delivery date. Products that are MTO are often unique and customized. The associated learning curve slope and other production parameters cannot be precisely estimated before production starts. In this paper, a dynamic and adaptive approach to estimate the effects of learning and to optimize next period production is developed. This approach offers a closed-loop solution through stochastic dynamic programming. Monthly production data are used to update the joint probability distributions of production parameters via Bayesian methods. Our approach is illustrated using historical earned-value data from the Black Hawk Helicopter Program. Managerial insights are obtained and discussed.     

基于学习曲线的ERP实施培训时间决策模型

通过分析ERP项目的特点,将学习曲线引入到ERP实施的研究中,基于学习曲线理论刻画企业的工作效率随项目时间变化的规律.在此基础上,将ERP项目抽象为前期的培训和后期的具体实施两个阶段,以项目实施周期和咨询成本最小为目标,建立ERP实施培训时间决策模型,通过培训时间控制并预测ERP项目的整体周期,进一步利用遗传算法对模型进行优化求解.为区分信息化基础不同的企业在项目初期工作效率的差异,在模型中加入初始工作效率这一变量,通过企业的信息化基础水平间接地测算企业的初始工作效率;结合行为导向的实证研究结果,在模型中限定培训时间的取值范围,解决多个近似解之间的取舍问题.案例研究表明,模型在实际应用中可以起到较好的优化和预测效果,能够为企业制定ERP项目计划和投资预算提供决策支持.     

START UP OR LEARNING CURVES–SOME NEW APPROACHES

The traditional start up or learning curve is presented in a somewhat different way to encourage increased application of the start up curve as a useful tool for management planning, control, and decision purposes. It is shown that the parameter estimation problem can be avoided in those situations in which the percent start up or learning is known or can be estimated. Several interesting approaches allowing simplified computations are presented for those common situations in which (a) the point at which the marginal cost per unit stabilizes is known, (b) the cost of the first unit is known, and (c) the percent start up for both (a) and (b) is known. The relationship between breakeven analysis and start up curves is shown, as well as the relationship between marginal cost-revenue analysis and start up curves.     

Supply Chain Learning and Spillovers in Vendor Managed Inventory

Despite extensive literature on the value of supply chain collaboration programs, little research has examined the issue from the perspective of organizational learning. Using a unique, operational level dataset, we empirically examine the learning curves through which performance improvements are realized under vendor managed inventory (VMI). Performance is measured at the downstream distributor locations by examining inventory levels after controlling for customer service performance (stockouts). We identify and assess three sources of learning: a supply chain dyad's self‐learning, learning spillovers from electronic data interchange (EDI), and learning spillovers from other supply chain dyads. We find that self‐learning, learning spillovers from EDI, and learning spillovers from other supply chain dyads, all have positive and significant impacts on a distributor's inventory performance. In addition, we find that self‐learning may exhibit a U‐shaped learning curve (i.e., performance first improves and then plateaus or declines). These findings suggest that the various learning experiences with VMI and EDI can lead to improved performance over time, but the path to improvement may be complex.     

Explaining the Diffusion of Knowledge Management: The Role of Fashion

This paper provides evidence on the emergence and diffusion of the discourse of knowledge management. A literature review of the knowledge management and learning organization literatures demonstrates the lack of learning from one discourse to another and major differences in the concerns and issues that they address. At the same time, evidence on the level of interest in each discourse shows a tendency towards a normal curve distribution. Analytically, these findings suggest that the widespread diffusion of knowledge management might be explained in terms of the management fashion model. However, further consideration of the professionally-differentiated appropriation of knowledge management concepts by the information systems and human resource communities suggests that the fashion model provides only a partial explanation for the observed diffusion of knowledge management.     

代价敏感的目标客户选择半监督集成模型研究

在现实的目标客户选择建模中,往往只能获取少量有类别标签的样本,而剩下的大量样本都无法获取类别标签。已有研究大都使用监督式建模研究范式,仅在少量有类别标签样本集上建模,很难取得令人满意的效果。为解决这一问题,本文引入半监督学习（semi-supervised learning,SSL）技术,将其与代价敏感学习（cost sensitive learning,CSL）和多分类器集成中的随机子空间（random subspace,RSS）方法相结合,提出了代价敏感的目标客户选择半监督集成模型（cost-sensitive semi-supervised ensemble model,CSSE）。该模型使用代价敏感的支持向量机（SVM）来解决目标客户选择建模中样本数据类别分布不平衡的问题,还能够同时使用有、无类别标签的客户样本来建模。进一步地,该模型利用RSS方法训练一系列基本分类模型,并通过集成得到最终的分类结果。在某保险公司目标客户选择数据集上进行实证分析,结果表明,与两种监督式集成模型、两种单一的半监督模型以及两种半监督集成模型相比,CSSE模型具有更好的目标客户选择性能。     

Can good leadership be learned through business games?

This article explores the use of business games to question whether they can be used to learn good leadership. The answer comes in three parts: yes, no and maybe. By studying the use of a game to teach change management in the light of adult learning theory, the article discusses how different kinds of learning processes emerge and contribute. On the basis of qualitative interviews and observation, a case study is used to explore how the particular, didactic orchestration of a game-based learning process, together with emergent, informal learning processes, contributes to leadership learning. As participants are immersed in formal, non-formal and informal learning processes, these processes contribute to improving leadership in practice in a variety of different ways. The article concludes that (1) yes, game-based learning does contribute to teaching operational leadership in a formal manner; (2) no, in and of itself, the formal learning of games has a limited impact on leadership learning, but in combination with other didactic activities, its contribution is profound and (3) maybe, since by tapping into the extensive knowledge of other participants, games inspire informal discussions on leadership topics that are not directly relevant to the game, but are in other ways beneficial.     

基于知识管理的敏捷组织学习二维度模型框架

如何在知识管理先进思想的基础上,充分利用信息技术来促进组织学习,是当前企业管 理的主要研究问题之一. 在剖析了两个具有代表性的组织学习模型基础上,提出了一种基于知 识管理的敏捷组织学习二维度模型框架. 在该模型框架中,充分考虑了组织学习的核心内容, 并围绕学习主体和学习工具的互动关系,对组织学习过程进行了构建. 文章不仅详细给出了该 模型框架的建立过程以及实施要素、原则和方法,而且通过一个案例具体说明了该模型框架在 企业项目建设中的实际意义和应用价值.     

具有最优学习率的RBF神经网络及其应用

传统固定学习率的RBF神经网络在金融时间序列预测方面已经有比较成功的应用,但网络学习率的选择问题却给传统RBF神经网络的使用带来了不便.利用梯度下降法及优化方法推导出了RBF神经网络的动态最优学习率并将其应用于网络学习算法,具有最优学习率的RBF神经网络能够在保证网络稳定学习的同时兼顾网络的收敛速度.为了检验具有动态最优学习率的RBF神经网络的预测效果,对沪深300指数波动率进行了预测实验.实验结果表明,具有动态最优学习率的RBF神经网络比传统的固定学习率的RBF神经网络有着更快的收敛速度,同时也避免了人为选定学习率的不便.     

Inductive,Evolutionary, and Neural Computing Techniques for Discrimination: A Comparative Study*

This paper provides a comparative study of machine learning techniques for two-group discrimination. Simulated data is used to examine how the different learning techniques perform with respect to certain data distribution characteristics. Both linear and nonlinear discrimination methods are considered. The data has been previously used in the comparative evaluation of a number of techniques and helps relate our findings across a range of discrimination techniques.     

Identifying and managing persisting tensions affecting strategic learning from projects

While research recognises that strategic learning from projects helps build organisation-wide knowledge capabilities, learning from projects still remains fragmented and is often prematurely discontinued. Although research demonstrated that persisting tensions affect project-based organising, little is known about specific tensions affecting strategic learning from projects. A single, in-depth case study in a multinational food processing and packaging corporation is used to explore paradoxical tensions affecting strategic learning from projects. The study uncovered three persisting learning tensions: project/organisational identity, short-/long-term-orientation and standardisation/flexibility. Seven strategies are proposed, supported by the empirical data, to respond to these tensions.     

Examining the Impact of Speed of Quality Improvement on Quality-Related Costs

The Fine (1986) quality-based learning curve model is extended to include the consideration of speed of quality improvement. The model demonstrates that under different circumstances rapid quality improvement effects are either beneficial or detrimental to improvement in quality-related costs. Hypotheses are developed from the analysis of this speed of quality improvement model. The hypotheses are tested in an automotive parts manufacturing company with five similar plants. Results show that with an increase in the speed of quality improvement, the rate of growth in prevention and appraisal costs decrease and the rate of growth in failure costs are unaffected. Rapid speed of quality improvement does yield lesser decreases in failure costs than slower, steadier improvement. However, rapid speed of quality improvement does not yield the predicted lesser decrease in prevention and appraisal costs than slower, steadier improvement. Rapid speed of quality improvement might or might not benefit the organization, perhaps an explanation for some Total Quality Management (TQM) failures. A more deliberate, learning organization is suggested from this research.     

‘Talking the talk’ – a discursive approach to evaluating management development

Much evaluation practice is guided by a theoretical anachronism. Systematized and formulaic approaches of chain reaction models of evaluation are based on conceptions of management which are over 50 years old. This study presents an evaluation of a management development intervention which used evidence of discursive shift in learners as an indicator of attitudinal and behavioural change brought about by the learning programme. The approach taken characterizes the learning event and the evaluation process as interventions which have the capacity to help learners to shape their identity and in which language is a mediator of learning. The results of the study, which applies both quantitative and qualitative techniques, are used to argue in favour of the deployment of a model of evaluation which has a focus on social and relational aspects of a managers' role.