Learning multicriteria classification models from examples: Decision rules in continuous space

The classification problem statement of multicriteria decision analysis is to model the classification of the alternatives/actions according to the decision maker's preferences. These models are based on outranking relations, utility functions or (linear) discriminant functions. Model parameters can be given explicitly or learnt from a preclassified set of alternatives/actions.In this paper we propose a novel approach, the Continuous Decision (CD) method, to learn parameters of a discriminant function, and we also introduce its extension, the Continuous Decision Tree (CDT) method, which describes the classification more accurately.The proposed methods are results of integration of Machine Learning methods in Decision Analysis. From a Machine Learning point of view, the CDT method can be considered as an extension of the C4.5 decision tree building algorithm that handles only numeric criteria but applies more complex tests in the inner nodes of the tree. For the sake of easier interpretation, the decision trees are transformed to rules.     

Decision under risk as a multicriteria problem

Most of the approaches to decision problems under uncertainty are based on decision paradigms, generally associated to an optimization process that leads to a final solution. For the Decision Maker, the basic decision is thus what paradigm to choose, the rest of the procedure being mainly technical.     

A multicriteria DSS for stock evaluation using fundamental analysis

The paper describes a multicriteria decision support system which aims at presenting an evaluation of the Athens Stock Exchange (ASE) stocks, on the basis of fundamental analysis. The system evaluates the stocks based on the method of fundamental analysis ratios, which is the most appropriate evaluation approach regarding investment decisions within a long term horizon. In addition to quantitative data deriving from fundamental analysis, the system uses qualitative data as well, in order to improve the reliability of the evaluation. The system introduced in this paper, utilises multicriteria analysis methodologies in order to rank the stocks by placing the best stock first and the worst last. Stock evaluation considers the specific characteristics of the potential investor, as well as his attitude towards undertaken risk. The final output of the system is four stock rankings which respond to four different criteria groups, depending on the type of accounting plan each listed company belongs to. The system incorporates a large volume of relevant information and operates in ‘real world conditions’ since its data are constantly updated. Finally, the system is intended for both institutional and private investors.     

Learning lexicographic orders

The purpose of this paper is to learn the order of criteria of lexicographic decision under various reasonable assumptions. We give a sample evaluation and an oracle based algorithm. In the worst case analysis we are dealing with the adversarial models. We show that if the distances of the samples are less than 4, then it is not learnable, but 4-distance samples are polynomial learnable.     

Cost-efficiency analysis of weapon system portfolios

Decisions about the acquisition and maintenance of military equipment serve to build long-term capabilities in preparation of military conflicts. Typically, these decisions involve large investments which need to be supported by adequate cost-efficiency analyses. Yet the cost-efficiency analysis of weapon systems involves several challenges: for example, it is necessary to account for the possible interactions among different weapon systems; the relevance of several impact criteria; and the variety of combat situations in which these systems may be used. In this paper, we develop a portfolio methodology where these challenges are addressed by evaluating the cost-efficiencies of entire portfolios consisting of individual weapon systems. Our methodology accounts for possible interactions among systems by synthesizing impact assessment results that are either generated by combat simulation models or elicited from experts. It also admits incomplete preference information about the relative importance of different impact criteria. This methodology guides decision making by identifying which combinations of weapon systems are efficient with respect to multiple evaluation criteria in different combat situations at different cost levels. It can also be extended to settings where multiple combat situations are addressed simultaneously. The methodology is generic and can therefore be applied also in civilian settings when portfolios of activities (such as mitigation of harmful environmental emissions) may exhibit interactions.     

The anticipative concept in warehouse optimization using simulation in an uncertain environment

The modern business environment is highly unpredictable. An anticipation approach in a real case study is presented to cope with such instability and minimize the total inventory cost without stock-outs occurring and inventory capacity being exceeded. The anticipation concept is performed using simulation models supported by inventory control algorithms on a selected sample of representative items. The inventory control algorithms include Silver–Meal, Part period balancing, Least-unit cost, and Fuzzy inventory control algorithm based on fuzzy stock-outs, highest inventory level and total cost. Transportation cost is explicitly defined as a discrete function of shipment size. The algorithms are tested on historic data. Simulation results are presented and the risk of accepting them as reliable is discussed. The process of simulation model implementation is briefly discussed to further validate the model and train order managers to use the simulation model in their order placement process.     

Fuzzy analytic hierarchy process: Fallacy of the popular methods

The fuzzy Analytic Hierarchy Process (fuzzy AHP) is a very popular decision making method and literally thousands of papers have been published about it. However, we find the basic logic of this approach has problems. From its methodology, the definition and operational rules of fuzzy numbers not only oppose the main logic of fuzzy set theory, but also oppose the basic principles of the AHP. In dealing with the outcomes, fuzzy AHP does not give a generally accepted method to rank fuzzy numbers and a way to check the validity of the results. Besides, we discuss the validity of the Analytic Hierarchy/Network Process (AHP/ANP) in complex and uncertain environments and find that fuzzy ANP is a false proposition because there is no fuzzy priority in the super matrix which provides the basis for the ANP. Although fuzzy AHP has been applied in many cases and cited hundreds of times, we hoped that those who use fuzzy AHP would understand the problems associated with this method.     

An integrative evaluation framework for intelligent decision support systems

Evaluation of the overall effectiveness of decision support systems (DSS) has been a research topic since the early 1980s. As artificial intelligence methods have been incorporated into systems to create intelligent decision support systems (IDSS), researchers have attempted to quantify the value of the additional capabilities. Despite the useful and relevant insights generated by previous research, existing evaluation methodologies offer only a fragmented and incomplete view of IDSS value and the contribution of its technical infrastructure. This paper proposes an integrative, multiple criteria IDSS evaluation framework through a model that links the decision value of an IDSS to both the outcome from, and process of, decision making and down to specific components of the IDSS. The proposed methodology provides the designer and developer specific guidance on the intelligent tools most useful for a specific user with a particular decision problem. The proposed framework is illustrated by evaluating an actual IDSS that coordinates management of urban infrastructures.     

Verbal ordinal classification with multicriteria decision aiding

Professionals in neuropsychology usually perform diagnoses of patients’ behaviour in a verbal rather than in a numerical form. This fact generates interest in decision support systems that process verbal data. It also motivates us to develop methods for the classification of such data. In this paper, we describe ways of aiding classification of a discrete set of objects, evaluated on set of criteria that may have verbal estimations, into ordered decision classes. In some situations, there is no explicit additional information available, while in others it is possible to order the criteria lexicographically. We consider both of these cases. The proposed Dichotomic Classification (DC) method is based on the principles of Verbal Decision Analysis (VDA). Verbal Decision Analysis methods are especially helpful when verbal data, in criteria values, are to be handled. When compared to the previously developed Verbal Decision Analysis classification methods, Dichotomic Classification method performs better on the same data sets and is able to cope with larger sizes of the object sets to be classified. We present an interactive classification procedure, estimate the effectiveness and computational complexity of the new method and compare it to one of the previously developed Verbal Decision Analysis methods. The developed and studied methods are implemented in the framework of a decision support system, and the results of testing on artificial sets of data are reported.     

Multiple-attribute decision support system based on fuzzy logic for performance assessment

This paper deals with the problem of assessing the performance of a set of production units, simultaneously considering different kinds of information, yielded by a Data Envelopment Analysis, a qualitative data analysis and an expert assessment. The tool for integrating heterogeneous data is a model that applies fuzzy logic to decision support systems. The results obtained are a holistic performance assessment of each unit of the set and a ranking order of the units.     

House selection via the internet by considering homebuyers’ risk attitudes with S-shaped utility functions

The widespread use of the Internet has significantly changed the behavior of homebuyers. Using online real estate agents, homebuyers can rapidly find some modern houses that meet their needs; however, most current online housing systems provide limit features. In particular, existing systems fail to consider homebuyers’ housing goals and risk attitudes. To increase effectiveness, online real estate agents should provide an efficient matching mechanism, personalized service and house ranking with the aim of increasing both buyers’ satisfaction and deal rate. An efficient online real estate agent should provide an easy way for homebuyers to find (rank) a suitable house (alternatives) with consideration of their different housing philosophies and risk attitudes. In order to comprehend these ambiguous housing goals and risk attitudes, it is also indispensable to determine a satisfaction level for each fuzzy goal and constraint.     

Handling effects of reinforced preference and counter-veto in credibility of outranking

The aim of this paper is to generalize the way of computing the credibility of outranking in a multiple criteria aggregation procedure, in view of taking into account two new effects called reinforced preference and counter-veto. These effects concern only those criteria for which, as soon as action a is “judged very strongly preferred” to action b, one wishes that the credibility of outranking of a over b is greater than that for the case where (all things equal elsewhere) the preference is not “judged very strong”. To achieve this goal, we propose two complementary ways. The first one involves a reinforced preference threshold which affects the concordance degree, and the second one involves a counter-veto threshold which affects the insertion of discordance degree in the calculation of the credibility of outranking. The introduction of these two new effects remains compatible with the handling of ordinal preference scales. The resulting new index of the credibility of outranking can be used, in particular, in ELECTRE methods.     

Modelling impacts of cropping systems: Demands and solutions for DEX methodology

Decision modelling of diverse groups of problems makes different requirements to the modelling methodologies and software. We present an actual decision problem and the required characteristics of corresponding decision models. The problem is from agronomy and addresses the ecological and economic impacts of cropping systems, with the focus on the differences between cropping systems with conventional crops and the ones with genetically modified crops. We describe the extensions of an existing DEX qualitative multi-attribute modelling methodology, which were made to cope with the challenges of the problem. The extensions address general hierarchical structures, probabilistic utility functions and numerical values of basic attributes. A new, freely available software tool called proDEX was implemented to support the extended methodology. In this paper we describe the problem of cropping system assessment, propose methodological extensions to DEX, and present the implementation of proDEX.     

Multi-criteria decision aid for the formulation of sustainable technological energy priorities using linguistic variables

Implementation of new and innovative energy technologies is a key mean towards a sustainable energy system. Currently, governments have to decide from an increasingly diverse mix of them, the ones which warrant support, including funding and other incentives for private sector efforts. However, appraising energy technologies in terms of their sustainability is a really complex task, considering the series of uncertainties and implications that have to be encountered so as to obtain realistic and transparent results. In this context, the main aim of this paper is to present a direct and flexible multi-criteria decision making approach, using linguistic variables, to assist policy makers in formulating sustainable technological energy priorities. Furthermore, its software realization will be applied to a number of technologies, in the context of the Greek Technology Foresight Programme, and the results will be presented and discussed.     

An evolutionary approach to construction of outranking models for multicriteria classification: The case of the ELECTRE TRI method

Outranking methods constitute an important class of multicriteria classification models. Often, however, their implementation is cumbersome, due to the large number of parameters that the decision maker must specify. Past studies tried to address this issue using linear and nonlinear programming, to elicit the necessary preferential information from assignment examples. In this study, an evolutionary approach, based on the differential evolution algorithm, is proposed in the context of the ELECTRE TRI method. Computational results are given to test the effectiveness of the methodology and the quality of the obtained models.     

Supporting triage of children with abdominal pain in the emergency room

The paper describes a methodology used for selecting the most relevant clinical features and for generating decision rules based on selected attributes from a medical data set with missing values. These rules will help emergency room (ER) medical personnel in triage (initial assessment) of children with abdominal pain. Presented approach is based on rough set theory extended with the ability of handling missing values and with the fuzzy measures allowing estimation of a value of information brought by particular attributes. The proposed methodology was applied for analyzing the data set containing records of patients with abdominal pain, collected in the emergency room of the cooperating hospital. Generated rules will be embedded into a computer decision support system that will be used in the emergency room. The system based on results of presented approach should allow improving of triage accuracy by the emergency room staff, and reducing management costs.     

Using a DEA-cross efficiency approach in public procurement tenders

The paper addresses the topic of supplier selection in public procurement. According to European directives, when tenders are awarded through the “Most Economically Advantageous Tender” (MEAT) criterion, the awarding committee has to decide the tender evaluation criteria of the presented bids in advance. The authors propose a decision making tool that is aimed at helping the awarding committee in this difficult task and, at the same time, maintaining a transparent procedure in accordance with governmental procurement regulations and requirements as well as guaranteeing fair and equal evaluation of all bids. In this regard, the decision problem of supplier selection is addressed by applying an extension of the DEA (data envelopment analysis) methodology. The cross-efficiency evaluation is used for selecting the best supplier among the eligible candidates. The proposed technique allows the evaluation of quantitative data related to vendor selection and keeps the transparency features requested by public procurement. In addition, all bids are equally assessed according to the same objectively defined weights without any subjective setting by the public officers. The effectiveness and efficiency of the approach is supported by a case study that pertains to the tender of an Italian public agency.     

Fuzzy set approach to the utility,preference relations,and aggregation operators

Score x = (x₁, … , x_n) describing an alternative α is modelled by means of a continuous quasi-convex fuzzy quantity μ_α = μ_x, thus allowing to compare alternatives (scores) by means of fuzzy ordering (comparison) methods. Applying some defuzzification method leads to the introduction of operators acting on scores. A special stress is put on the Mean of Maxima defuzzification method allowing to introduce several averaging aggregation operators. Moreover, our approach allows to introduce weights into above mentioned aggregation, even in the non-anonymous (non-symmetric) case. Finally, Ordered Weighted Aggregation Operators (OWAO) are introduced, generalizing the standard OWA operators.