Influence of model management systems on decision making: empirical evidence and implications

Model management (MM) regards decision models as an important organisational resource deserving prudent management. Despite the remarkable volume of model management literature compiled over the past twenty-odd years, very little is known about how decision makers actually benefit from employing model management systems (MMS). In this paper, we report findings from an experiment designed to verify the idea that the adequacy of modeling support provided by a MMS influences the decision maker's problem solving performance and behaviour. We show that the decision makers who receive adequate modelling support from MMS outperform those without such support. Also, we provide empirical evidence that the MMS help turn the decision makers' perception of problem solving from a number crunching task into development of solution strategies, consequently changing their decision making behaviour.     

MEDSS: A Multi-Agent Environmental Decision Support System framework for model management

Model Management Systems (MMS) have become increasingly important in handling complicated problems in Decision Support Systems (DSS). The primary goal of MMS is to facilitate the development and the utilization of quantitative models to improve decision performance. Much current research focuses on model construction. Where early research used deductive reasoning approaches to construct new models, more recent efforts use inductive reasoning mechanisms. Both approaches have their drawbacks. Deductive reasoning methods require a strong domain theory (which may not exist or may be too complex to apply) and ignore previous solving experience. Inductive reasoning methods can take advantage of precedents or prototypical cases, but do not employ domain knowledge. Both methods are limited in learning capacity. This study proposes a Multi-Agent Environmental Decision Support System, which integrates an Inductive Reasoning Agent, and an Environmental Learning Agent to perform new model formation and problem solving. New models can be generated by the coordination of both the Inductive Agent and the Deductive Agent. At the same time, a model repair process is undertaken by the Environmental Learning Agent when the prediction resulting from existing knowledge fails.     

THE USE OF TELESCOPING SPATIAL SCALES TO CAPTURE INSHORE TO SLOPE DYNAMICS IN MARINE ECOSYSTEM MODELING

Abstract Ecosystem processes function at many scales, and capturing these processes is a challenge for ecosystem models. Nevertheless, it is a necessary step for considering many management issues pertaining to shelf and coastal systems. In this paper, we explore one method of modeling large areas with a focus at a range of scales. We develop an ecosystem model that can be used for strategic management decision support by modeling the waters off southeastern Australia using a polygon telescoping approach, which incorporates fine‐scale detail at the coastal zone, increasing in scale to a very coarse scale in the offshore areas. This telescoping technique is a useful tool for incorporating a wide range of habitats at different scales into a single model.     

Designing risk management support systems

The major movements emerging in the field of risk analysis over the past decades are reviewed with special emphasis placed on their treatment of non-technical issues and their relevance to risk management needs. With this background, the requirements for an effective system to support risk management activities are discussed and a user-oriented design process outlined. The Ispra Risk Management Support (IRIMS) System is briefly described.Invited for presentation at the 12th Symposium on Operations Research, University of Passau, Federal Republic of Germany, September 9–12, 1987.     

盈余管理计量模型效力的实证研究

本文以随机抽取的上海和深圳证券交易所的1000家被并购上市公司为样本,研究了盈余管理计量模型的检验效力问题。本文通过三种验证各模型效力的实证方法比较了6种盈余管理的计量模型。研究发现:第一类误差的方法认为Healy模型、修正的琼斯模型和前瞻性修正的琼斯模型能有效地检验出盈余管理行为:预测误差标准方法则认为前瞻性修正的琼斯模型是最优的模型:而根据对操纵性应计部分的短期效应的实证则认为Healy模型、修正的琼斯模型、KS模型和前瞻性修正的琼斯模型是较优的计量模型。在综合考虑各种实证结果和中国资本市场实际情况的基础上,认为前瞻性修正的琼斯模型在中国证券市场检验盈余管理行为的效力最强。     

An architecture for knowledge evolution in organisations

Methodology for Knowledge System Management (MKSM) is a methodology, based on modelling techniques, to support knowledge capitalisation and management. This kind of approach faces a new challenge: knowledge evolution. Indeed, as knowledge of the organisation evolves, it seems necessary to have models and supporting tools to represent this evolution. Biologic evolution theories can offer the basis of a dynamic theory of Knowledge System evolution. We especially propose to adapt Lamarckism's principles and the Cladistic classification to support such a modelling and propose a conceptual architecture for managing knowledge evolution.     

An agent-based investigation into the new product development capability

The capability to bring products to market which comply with quality, cost and development time goals is vital to the survival of firms in a competitve environment. New product development comprises knowledge creation and search and can be organized in different ways. In this paper, we study the performance of several alternative organizational models for new product development using a model of distributed, self-adapting (learning) agents. The agents (a marketing and a production agent) are modelled via neural networks. The artificial new product development process analyzed starts with learning on the basis of an initial set of production and marketing data about possible products and their evaluation. Subsequently, in each step of the process, the agents search for a better product with their current models of the environment and, then, refine their representations based on additional prototypes generated (new learning data). Within this framework, we investigate the influence of different types of new product search methods and generating prototypes/learning according to the performance of individual agents and the organization as a whole. In particular, sequential, team-based Trial & Error and House of Quality guided search are combined with prototype sampling methods of different intensity and breadth; also, the complexity of the agents (number of hidden units) is varied. It turns out that both the knowledge base and the search procedure have a significant impact on the agents' generalization ability and success in new product development. Andreas Mild was born in Vienna, Austria, in 1973. He studied business administration in Vienna, in 2000 he received his Ph.D. from the Vienna University of Economics and Business Administration (WU). Since 2003 he is associated professor at the WU. He has been guest professor in Frankfurt, Germany, Sydney, Australia and Bangkok, Thailand. Previous research appeared in Journals such as MIS Quarterly, Management Science and Marketing Science. His research interests currently include agent-based models, new product development and recommender systems. Alfred Taudes was born in Vienna, Austria, in 1959. He studied business administration and management information systems (MIS) in Vienna (doctorate 1984), in 1991 he received his Ph.D. from the Vienna University of Economics and Business Administration (WU). He was assistant professor at the WU (1986–1991) and professor for MIS at the German Universities of Augsburg (1991), Münster (1991/92) and Essen (1992/93). Since 1993, he has been professor for MIS at the WU and Head of the Department for Production Management. Since 2000, Dr. Taudes has been speaker for the Special Research Area SFB # 010 (Adaptive Information Systems and Modelling in Economics and Management Science). His research interests currently include agent-based models of industry structures, management of innovation, technology management and business strategy.     

On substitutability and complementarity in discrete choice models

In this paper, we propose the concepts of substitutability and complementarity in discrete choice models. These concepts concern whether the choice probability of one alternative in a choice model increases or decreases with the utility of another alternative, and they play important roles in capturing certain practical choice patterns, such as the halo effect. We study conditions on discrete choice models that will lead to substitutability and complementarity. We also present ways of constructing choice models that exhibit complementary property.     

Economic design of integrated model of control chart and maintenance management

There are two key tools for the control of production processes — Statistical Process Control (SPC) and Maintenance Management (MM), which are traditionally separated (both in science and in business practice), even though their goals overlap a great deal. Their common goal is to achieve optimal product quality, little downtime and cost reduction by controlling variances in the process. Since single or separated parallel applications may not be fully effective, this paper discusses the integration of statistical process control and maintenance, and provides an integrated model of Control Chart (CC) and MM. A mathematical model is given to analyze the cost of the integrated model and the grid-search approach is used to find the optimal values of policy variables (n,h,L,k) that minimize hourly cost. Finally, a numerical experiment is conducted to investigate the effects of cost parameters on the solution of the design.     

Development planning based on interoperable agent driven simulation

The paper presents the potential of using interoperable agent driven simulation to support development planning; indeed the use of simulation represents a strong benefit to improve planning of infrastructures and plants devoted to disaster relief, civil protection and/or support to country development; the paper describes models used to face these challenges and last updates in population modeling for these applications. The proposed models include population characteristics, need as well as their social networks. In humanitarian support operations and country reconstruction there is a huge potential to use simulators; the paper describes how these models should be designed to support training as well operational planning. The models should be able to consider the impacts of contingencies as well as to guarantee the quick responsiveness requirements for humanitarian crisis management. The authors propose a simulator to be shared and used among Armed Forces and Civil Agencies for addressing Crisis Management, Humanitarian Missions, Country Reconstruction and Development considering joint operations (i.e. Civil Military Cooperation); indeed the paper outlines the importance of training people devoted to guarantee interoperability among civil organization and military units in this sector. The paper describes the models based on interoperable simulation as well as the agents driving the entities during the simulation to create quickly complex scenarios able to consider the impact on population and communities of the different actions by including human behavioral models. The proposed approach guarantees interoperability among different simulators within an HLA (High Level Architecture) federation in order to recreate crisis scenarios combining detailed simulation of multiple factors. The proposed approach is verified and validated by proposing an experimental analysis where it is evaluated a set of construction projects (i.e. digging wells) in a stabilization area and their effectiveness both in terms of direct result (i.e. water availability) as well as of population consensus and disaster relief (i.e. stress mitigation, trustiness respect supporting players).     

Capturing and prioritizing students’ requirements for course design by embedding Fuzzy-AHP and linear programming in QFD

Customer requirements play a vital and important role in the design of products and services. Quality Function Deployment (QFD) is a popular, widely used method that helps translate customer requirements into design specifications. Thus, the foundation for a successful QFD implementation lies in the accurate capturing and prioritization of these requirements. This paper proposes and tests the use of an alternative framework for prioritizing students’ requirements within QFD. More specifically, Fuzzy Analytic Hierarchy Process (Fuzzy-AHP) and the linear programming method (LP-GW-AHP) based on Data Envelopment Analysis (DEA) are embedded into QFD (QFD-LP-GW-Fuzzy AHP) in order to account for inherent subjectivity of human judgements. The effectiveness of the proposed framework is assessed in capturing and prioritizing students’ requirements regarding courses’ learning outcomes within the process of an academic course design. Sensitivity analysis evaluates the robustness of the prioritization solution and implications for course design specifications are discussed.     

Computer-assisted Modelling and Analysis of Linear Programming Problems: Towards a Unified Framework

A framework for model formulation and analysis to support operationsand management of large-scale linear programs is developed fromthe combined capabilities of CAMPS and ANALYZE. Both the systemsare reviewed briefly and the interface which integrates thetwo systems is then described. The model formulation, matrixgeneration, and model management capability of CAMPS and thecomplementary model and solution analysis capability of ANALYZEare presented within a unified framework. Relevant generic functionsare highlighted, and an example is presented in detail to illustratethe level of integration achieved in the current prototype system.Some new results on discourse models and model management supportare given in a framework designed to move toward an ‘intelligent’system for linear programming modelling and analysis.     

考虑新增样本的中国上市公司马尔科夫分析

使用吸收马尔科夫链对中国上市公司的未来经营状况进行了分析和预测.在前人研究的基础上,本文将相应的分析进行了扩展,考虑了新增样本情况下的上市公司整体经营状况.实证研究表明,考虑新增样本情况下的上市公司整体经营状况与不考虑此因素的结果稍有不同,而这正反映了新增样本的影响.     

A FULL ANNUAL CYCLE MODELING FRAMEWORK FOR AMERICAN BLACK DUCKS

American black ducks (Anas rubripes) are a harvested, international migratory waterfowl species in eastern North America. Despite an extended period of restrictive harvest regulations, the black duck population is still below the population goal identified in the North American Waterfowl Management Plan (NAWMP). It has been hypothesized that density‐dependent factors restrict population growth in the black duck population and that habitat management (increases, improvements, etc.) may be a key component of growing black duck populations and reaching the prescribed NAWMP population goal. Using banding data from 1951 to 2011 and breeding population survey data from 1990 to 2014, we developed a full annual cycle population model for the American black duck. This model uses the seven management units as set by the Black Duck Joint Venture, allows movement into and out of each unit during each season, and models survival and fecundity for each region separately. We compare model population trajectories with observed population data and abundance estimates from the breeding season counts to show the accuracy of this full annual cycle model. With this model, we then show how to simulate the effects of habitat management on the continental black duck population.     

L.P.L. a fuzzy programming language: 1. Syntactic aspects

The L.P.L. language (Linguistic oriented Programming Language) that we shall present in this paper and in a forthcoming one is a new language aiming at an implementation of the concepts of the fuzzy sets theory. In the first paper, after an introduction to the basic concepts. we shall describe the syntactic aspects of L.P.L., i.e., the data declarations structure, the statements structure, the general structure of L.P.L. models. The second paper will deal first with the basic semantic and the semantics of logical expressions, then the semantics of basic statements as well as the semantics of the control structure of programs.     

The Theory of SML Schema-Directed Query

SML is a modeling language for the structured modeling framework, which represents the semantics as well as the mathematical structure of a model. This paper uses an SML approach to improve the object based universal relation data model. By this approach, both the relational structure of a database and the objects in relations are automatically derived by the associated SML schema. The interpretation part of an SML schema allows users to easily learn the meanings of the data before performing universal relation queries; the queries are then computed by using the automatically derived objects. With a goal of making queries simpler, this paper presents theories, table naming conventions, a confirmation approach, and a unified example illustrating many different concepts. It helps lay the foundation for the eventual development of a remarkably easy user interface for ad hoc query in computer-based modeling systems. We are hopeful that the results may in the future contribute to real applications in databases as well as in management science/operations research.     

Introduction to the Special Issue on Discrete Event Models of Continuous Systems

Automatic control systems rely on models to predict the near future and identification algorithms to adapt the models to changing process behaviour. The traditionally highly complex models of the activated sludge process developed for scientific purposes cannot be identified from on-line measurements and are not suited for process control purposes in their present form. Model decoupling based on the different time scales of the dynamic processes is one possible way of attacking this problem. It allows the implementation of more simple and realistically applicable controllers in combination with predictions based on simplified models in hierarchical control structure. This paper discusses these concepts and presents a reduced order model describing carbonaceous removal, nitrification, and denitrification in a medium time scale (several hours/days). The model parameters are identifiable from available on-line measurements and the dynamic behaviour is verified against computer simulations of the IAWQ activated sludge model no. 1.     

The ANALYZE rulebase for supporting LP analysis

This paper describes how to design rules to support linear programming analysis in three functional categories: postoptimal sensitivity, debugging, and model management. The ANALYZE system is used to illustrate the behavior of the rules with a variety of examples. Postoptimal sensitivity analysis answers not only the paradigmWhat if …? question, but also the more frequently askedWhy …? question. The latter is static, asking why some solution value is what it is, or why it is not something else. The former is dynamic, asking how the solution changes if some element is changed. Debugging can mean a variety of things; here the focus is on diagnosing an infeasible instance. Model management includes documentation, verification, and validation. Rules are illustrated to provide support in each of these related functions, including some that require reasoning about the linear program's structure. Another model management function is to conduct a periodic review, with one of the goals being to simplify the model, if possible. The last illustration is how to test new rule files, where there is a variety of ways to communicate a result to someone who is not expert in linear programming.     

A PREDICTIVE MODEL TO INFORM ADAPTIVE MANAGEMENT OF DOUBLE‐CRESTED CORMORANTS AND FISHERIES IN MICHIGAN

The proliferation of double‐crested cormorants (DCCOs; Phalacrocorax auritus) in North America has raised concerns over their potential negative impacts on game, cultured and forage fishes, island and terrestrial resources, and other colonial water birds, leading to increased public demands to reduce their abundance. By combining fish surplus production and bird functional feeding response models, we developed a deterministic predictive model representing bird–fish interactions to inform an adaptive management process for the control of DCCOs in multiple colonies in Michigan. Comparisons of model predictions with observations of changes in DCCO numbers under management measures implemented from 2004 to 2012 suggested that our relatively simple model was able to accurately reconstruct past DCCO population dynamics. These comparisons helped discriminate among alternative parameterizations of demographic processes that were poorly known, especially site fidelity. Using sensitivity analysis, we also identified remaining critical uncertainties (mainly in the spatial distributions of fish vs. DCCO feeding areas) that can be used to prioritize future research and monitoring needs. Model forecasts suggested that continuation of existing control efforts would be sufficient to achieve long‐term DCCO control targets in Michigan and that DCCO control may be necessary to achieve management goals for some DCCO‐impacted fisheries in the state. Finally, our model can be extended by accounting for parametric or ecological uncertainty and including more complex assumptions on DCCO–fish interactions as part of the adaptive management process.