Practical Decision Support System Research Articles

Guest editorial: a special issue on optimization techniques for business intelligence systems

Business Intelligence refers to computer-based technologies, methodologies, tools, and systems for collecting, integrating, analyzing, and presenting large volumes of information to enable better business decision making. Increasingly, as businesses becomemore automated, data-driven, and real-time, business intelligence architecture evolves to support both strategic and operational levels of business decision making, which requires more advanced techniques and technical support. Aiming at better supporting business decision-making, business intelligence system is naturally a decision support system (DSS) in practice. Among many core techniques of DSS, an optimization-based DSS whose various optimizationmethodologies and techniques are applied, aims to arrive at the best ormost satisfactory decision out of a multitude of possible alternatives for business decision problems. Decision making has become more complicated and difficult in the current rapidly changing business environment. In recent years, both typical decision optimization such as multi-criteria decision making, and new decision intelligence techniques such as particle swarm optimization, have had unimaginable improvements. Particularly, through integrated with knowledge engineering and computational intelligence approaches, hybrid methodologies have been developed to deal with complex, uncertain, and un-structured decision optimization problems in businesses. This Special Issue provides an updated overview of the research field in line with optimization techniques for business intelligence systems. It includes ten papers to present recent developments in methodologies, techniques and applications in optimization techniques for business intelligence systems frommultiple aspects. The Special Issue covers three sub-topics: Intelligent multi-criteria optimization systems (papers 1–3); Optimization methodologies for business intelligence applications (papers 4–7); and Hybrid optimization algorithms by integration with computational intelligence and machine learning techniques (papers 8–10).

A support vector machine for decision support in melanoma recognition

The early diagnosis of melanoma is critical to achieving reduced mortality and increased survival. Although clinical examination is currently the method of choice for melanocytic lesion assessment, difficulties may arise in the diagnosis of atypical lesions. From both the naked eye and dermoscopic perspective, dysplastic naevi often exhibit a prominent heterogeneity of structure that renders their clinical distinction from melanoma difficult. To address these problems in diagnosis, there exists a heightened interest among researchers regarding the utility of machine learning techniques in computerised image analysis. Here we report on the utility, for dermatologists, of support vector machine (SVM) technology in melanoma diagnosis, using an archive of 199 digital dermoscopic images of excised atypical melanocytic lesions. Our best validation models achieved an average sensitivity and specificity for melanoma diagnosis of 0.86 and 0.72, respectively. Applying the best model to our test set yielded a sensitivity of 0.89, a diagnostic odds ratio of 14.09 and an area under the receiver operated characteristic curve (AUC) of 0.76. Advantages of the procedure implemented are the simplicity of feature extraction and the computationally cheap and efficient nature of SVMs. The derived model generalises well with outcomes that compare favourably with competing algorithms and expert assessment. In line with the concept of the utility of decision support systems in clinical practice, we propose that to reduce the risk of missed melanomas, both the dermatologists' assessment and the SVM diagnosis be incorporated into the clinical decision-making process.

Scheduling production for a sawmill: A comparison of a mathematical model versus a heuristic

Sawmill production scheduling is complex. It involves determining which logs to process taking into account the diameter, length, and grade of each log, on one hand, and the finished products that are needed to fulfill the orders, on the other. The cutting pattern determines which products are generated and also the yield, which is how much of the volume of the log ends as finished products. We used two approaches to solve the problem: a mathematical model, which calculates the volume and the schedule of various types of logs, and a heuristic that solves the problem by systematically applying the criteria used by a sawmill programmer in southern Chile. We compare the results of using both approaches under various scenarios in a 6 week planning horizon, with the mathematical model showing a superior performance in almost all instances except two, were both found the optimal solution. The proposed mathematical model can be solved in a relatively short time, which makes it a suitable basis of a practical optimization-based decision support system.

A matrix approach to status quo analysis in the graph model for conflict resolution

An algebraic method is developed to carry out status quo analysis within the framework of the graph model for conflict resolution. As a form of post-stability analysis, status quo analysis aims at confirming that possible equilibria, or states stable for all decision-makers, are in fact reachable from the status quo or any other initial state. Although pseudo-codes for status quo analysis have been developed, they have never been implemented within a practical decision support system. The novel matrix approach to status quo analysis designed here is convenient for computer implementation and easy to employ, as is illustrated by an application to a real-world conflict case. Moveover, the proposed explicit matrix approach reveals an inherent link between status quo analysis and the traditional stability analysis and, hence, provides the possibility of establishing an integrated paradigm for stability and status quo analyses.

Nurses’ use of computerised clinical decision support systems: a case site analysis

To explore how nurses use computerised clinical decision support systems in clinical practice and the factors that influence use. There is limited evidence for the benefits of computerised clinical decision support systems in nursing, with the majority of existing research focusing on nurses' use of decision support for telephone triage. Research has suggested that several factors including nurses' experience, features of the technology system and organisational factors may influence how decision support is used in practice. A multiple case site study. Four case sites were purposively selected to provide variation in staff experience, technology used and decisions supported by the technology. Data were collected in each case site using non-participant observation of nurse/patient consultations (n = 115) and interviews with nurses (n = 55). Data were analysed using thematic content analysis. Computerised decision support systems were used in a variety of ways by nurses, including recording information, monitoring patients' progress and confirming decisions that had already been made. Nurses' experience with the decision and the technology affected how they used a decision support system and whether or not they over-rode recommendations made by the system. The ability of nurses to adapt the technology also affected its use. How nurses use computerised decision support appears to be the result of an interaction between a nurses' experience and their ability to adapt the technology to 'fit' with local clinical practice. One of the stated aims of introducing computerised decision support systems to assist nursing practice is to reduce variation and/or the number of errors associated with clinical practice. The study found unanticipated uses in such systems such as the routine over-riding of recommendations which could lead to an increase rather than a decrease in variation or errors.

Bidding process and integrated fuzzy model for global sourcing based on customer preferences

Due to the increasing competition of globalisation and fast technological improvements, world markets demand companies to have quality and professional supply chain partners. This can only be achieved by employing potential bidding process for global sourcing. In this paper, we proposed a generic bidding process with respect to customer preferences for global sourcing. We develop an integrated fuzzy model leveraging fuzzy analytic hierarchy process and data envelopment analysis based on the model of Parthiban et al. (2008) for evaluation and selection of supply chain partner for global sourcing. The integrated fuzzy model is applied and illustrated for a real-life case study. A practical computer-based decision support system is developed for the model to help managers make better decisions under fuzzy circumstances.

A fuzzy AHP approach to personnel selection problem

Due to the increasing competition of globalization and fast technological improvements, world markets demand companies to have quality and professional human resources. This can only be achieved by employing potentially adequate personnel. In this paper, we proposed a personnel selection system based on Fuzzy Analytic Hierarchy Process (FAHP). The FAHP is applied to evaluate the best adequate personnel dealing with the rating of both qualitative and quantitative criteria. The result obtained by FAHP is compared with results produced by Yager's weighted goals method. In addition to above-mentioned methods, a practical computer-based decision support system is introduced to provide more information and help manager make better decisions under fuzzy circumstances.

Optimization methods applied for sustainable management of water-scarce basins

This paper aims to present the information-based technologies used for water management in a test basin within the scope of the EU-sponsored SMART and OPTIMA projects. The study uses a dynamic simulation model with its associated databases and a water resources planning and optimization system, established through a web-based client–server implementation to support distributed use and easy access for multi-criteria optimization and decision support. The results clearly illustrate how a consistent and well-integrated set of advanced but practical Decision Support System (DSS) tools can be used for efficient “optimal” water management strategies and policies of use, designed for a participatory public decision-making process.

Implementing a clinical decision-support system in practice: A qualitative analysis of influencing attitudes and characteristics among general practitioners

In order to obtain a better understanding of barriers to the implementation of clinical decision-support systems (CDSSs) in primary health care, we explored general practitioners' (GPs) handling of a CDSS during the implementation process. An Internet-based application for the management of chronic heart failure was used that was an adaptation of established clinical guidelines for computer use. The whole implementation process was followed closely, using a combination of different methods for data collection: repeated interviews with the five participating GPs, observations of patient visits, patient interviews, and detection of usage. We analysed the data using qualitative content analysis. The results showed that GPs' attitudes and characteristics constituted different profiles that seemed to be associated with the degree of acceptance of the CDSS. Those profiles were related to conceptions about the GPs' professional role and their attitudes towards the computer's function in disease management and in decision-making. Some additional barriers were insufficient level of computer skills and time constraints in everyday work. These findings can help us identify groups of GPs with definable needs during the implementation of a CDSS and make it easier to meet those needs with individually tailored interventions.

What middleware for network centric operations?

The main intent of this paper is to address the issue of middleware in network centric operations. To this end, we characterize a set of Information Technology capabilities that such a middleware should implement. Afterwards, we will discuss the design and architectural aspects of these capabilities. This will lead us to an efficient and practical decision support system that we call a digital cockpit. The latter is essentially a multi-tier IT platform that provides a plethora of services such as: data and service integration, monitoring, analysis, and process optimization. Moreover, the platform uses advanced display mechanisms to render the information in a structured and navigational representation that offers the possibility to drill down into the details. A significant subgoal of the paper is to discuss the quality attributes of such an NCO middleware. Finally, we present the results of an implementation of the aforesaid platform architecture.

Prospective Budgeting for Home Care

To provide a model for prospective budgeting for home care that is plausible, coherent, flexible, and sufficiently tractable that it can serve as a template for practical decision making and to clarify what would be the data requirements and statistical framework to calibrate the model. Methods used are standard risk-neutral expected value theory, cost benefit analysis, and the conditional logistic probability model. A simple but effective prospective budgeting model that provides analytic scaffolding for a practical decision support system for home care case managers, consultants, and program evaluators that can improve program equity, efficiency, and effectiveness. The author criticizes the well-known Titration Budgeting Model of Weissert, Chernew, and Hirth in terms of its logical and operational problems but then goes on to develop a framework within which the goals of the titration model can be met and home care resources can be more efficiently allocated.

Electronic decision support for procurement management: evidence on whether computers can make better procurement decisions

We analyse how well purchasing managers are able to judge the likelihood of problems for a given purchasing transaction. The literature on clinical versus statistical prediction suggests that humans in general, including purchasing managers, are often outperformed by relatively simple statistical formulas for such kinds of tasks. Based on a vignette experiment of real purchasing transactions, we compare the performance of purchasing managers with freshmen students and with a statistical formula based on a cross-validated sample. The results show that the formula outperforms the humans, and that experienced purchasing managers do not outperform freshmen students. We conclude that it would make sense to use decision support systems in the daily practice of purchase management so that humans can devote their time to what they are good at, while being guided by statistical software that takes care of multi-dimensional decisions in noisy environments.

A bilevel modelling approach to pricing and fare optimisation in the airline industry

The airline revenue management problem can be decomposed into four distinct but related sub-problems that are usually treated separately: demand forecasting, overbooking, capacity allocation and pricing. In recent decades, much interest has been devoted to overbooking and capacity allocation issues and, today, most major airlines rely on computerised tools to deal with these two sub-problems. Pricing, however, has received less attention, which can be explained by the technical and theoretical difficulties inherent in the implementation of a practical pricing decision support system. This paper presents a new modelling approach which allows for the joint solution of the capacity allocation and pricing sub-problems faced by a major North American airline. Using pre-defined booking limits, the resulting model can also be applied in a ‘pure’ pricing context. This approach is based on the bilevel programming paradigm, a special case of hierarchical mathematical optimisation. This modelling technique makes it possible to take into account matters such as customer segmentation, behaviour with regard to fares and other product attributes, and the interactions induced by overlapping routes, which are typical of modern air transportation networks.

Risk Stratification and Prevention in Chronic Coronary Artery Disease: Use of a Novel Prognostic and Computer-Based Clinical Decision Support System in a Large Primary Managed-Care Group Practice

Current therapy for chronic coronary artery disease (CAD) is extremely variable and under-utilizes preventive measures that have demonstrated benefit. Therefore, we sought to determine whether a novel hybrid prognostic and computer-based clinical decision support system (ohms|cad®; Interactive Heart Management Corporation, Laurence Harbor, New Jersey), that provides the treating physician with real-time, patient-specific recommendations, could improve outcomes, decrease utilization of services, and lower the costs of CAD. This was an observational study in a managed-care group practice with 39,565 members. A subpopulation of 1,628 patients participated in the study after being identified as being at increased risk for CAD. Over the 21 months after introduction of ohms|cad, utilization of cardiovascular services, coronary disease outcomes, and costs were compared with a baseline year by computing the relative risks of discrete CAD-related events. The use of ohms|cad significantly reduced adverse CAD outcomes in the entire population of members. Acute myocardial infarctions were reduced by 30% [relative risk (RR), 0.70; 95% confidence interval (CI), 0.59-0.81] and hospitalizations for angina or suspected infarction by 32% (RR, 0.68; 95% CI, 0.62-0.75). Stress tests decreased by 62% (RR, 0.38; 95% CI, 0.35-0.40), cardiac catheterizations by 20% (RR, 0.80; 95% CI, 0.71-0.89), and coronary angioplasties by 22% (RR, 0.78; 95% CI, 0.62-0.96). The rate of coronary artery bypass grafting was unchanged (RR, 1.06; 95% CI, 0.88-1.26). In addition, there were significant improvements in ambulatory ischemia, most monitored risk factors, and use of aspirin. Net CAD-associated costs in the population were reduced by 17%. Thus, an automated system that provides patient-specific prognostic information and clinical decision support for the treatment of chronic CAD can help increase the use of preventive strategies, and lower the morbidity and cost of chronic CAD.

Computer decision support systems in general practice

The majority of healthcare contacts in the United Kingdom occur in primary care. The development of structured knowledge-based systems is being encouraged to assist doctors in diagnosis (Shortliffe et al., Medical Informatics, Addison-Wesley, Workingham, 1990). These systems may improve the quality of medical care (Delaney et al., Br. Med. J. 319 (1999) 1281). Benefits can occur in disease prevention with an improvement in immunisation rates by up to 18% and other preventive tasks by up to 50% (Sullivan and Mitchell, Br. Med. J. 311 (1995) 848–852). Other benefits have been seen with disease management. More appropriate dosing has been seen in patients taking drugs with a narrow therapeutic range (Gonzalez et al., Computer-assisted optimisation of aminophylline therapy in the emergency department, Am. J. Emergency Med. 7 (1989) 395–401) and improvements have been shown in the management of chronic physical illness where there is a clear consensus regarding treatment protocols (Ryan et al., Health Care Computing 1 (1991) 332; Carson et al., Clinical decision support, systems methodology, and telemedicine: their role in the management of chronic disease, IEEE Trans. Inf. Technol. Biomed. 2 (1998) 80–88). However, the consultation time is lengthened by their use and there is no appreciable impact on patient satisfaction (Sullivan and Mitchell, Br. Med. J. 311 (1995) 848–852). Both practitioners’ conduct and the disclosure of information by the patient are adversely affected (Greatbatch et al., Fam. Practi. 12 (1995) 32–36), and many patients are concerned about confidentiality (Ridsdale and Hudd, Scand. J. Prim. Health Care 15 (1997) 180–183). Further, potential difficulties exist because of the wide variety of undifferentiated problems that are present in general practice. This makes evaluating the impact of computers in consultations particularly important (Feinstein, Lancet ii (1972) 421–423). These systems could potentially improve the quality of decisions made in the primary care consultation. However, considerable work is needed to ensure that their introduction is not detrimental to the quality of the relationship between the doctor and patient in the consultation, and to make the systems adaptable to local priorities. The systems need to be acceptable to both professionals and patients.

Decision support for the scheduling of subcontract manufacture

Presents work that addresses the issue of decision support for the operational management of subcontract manufacture. Theoretical study has been combined with empirical research and practical industrial investigation with regard to distributed manufacturing systems which incorporate subcontract manufacturing arrangements. There has been a particular focus on the study of resource planning and scheduling for subcontract manufacture in SMEs in the UK electronics assembly industry. This work led to the analysis, modelling, implementation and test of an object‐oriented advisory system to assist with scheduling for this domain which demonstrated the utility of a proposed concept of captivity‐based scheduling. Contemporary research in this area and existing commercial decision support solutions for manufacturing planning, scheduling and control in SMEs have been explored. Concludes that current commercial software systems for subcontract manufacture are underdeveloped. Further argues that software development tools and platforms are increasingly available to facilitate the creation of practical decision support systems for distributed organizational forms of manufacture.

Distributed Simulation of Freeway Traffic Flow Using Personal Computers

This article presents a personal computer (PC)-based distributed simulation methodology applicable to real-time estimation of time-variant parameters in freeway traffic flow. The method uses the interprocess communication procedures currently available under Windows NT and adopts 2 types of distributed computing architectures. The proposed method, based on macroscopic traffic models developed and tested in prior studies, first divides a given freeway into a set of subsections depending on the number of available processors or PCs. At every time step, each processor simulates its own subsection and exchanges the internal boundary data with another processor that simulates the adjacent subsection. The performance of the parallel algorithm was compared with that of a single-processor-based sequential simulation using 2 different types of hardware configurations: a network of 2 PCs and a dual-processor PC. Test results with the 2-PC network parallel system showed a speedup of 1.94 over the sequential simulation with a single PC for the 1-hour simulation of a 20-mile freeway section. The speedup for 1-hour simulation of an 80-mile section was increased to 1.96 with 2 processors, which is close to the ideal speedup. Performance tests with the dual-processor PC resulted in consistent, but not significantly faster, execution than those with the 2-PC network, indicating the feasibility of developing a practical on-line decision support system with a network of PCs for freeway operations.

A confident decision support system for interpreting electrocardiograms.

Computer-aided interpretation of electrocardiograms (ECGs) is widespread but many physicians hesitate to rely on the computer, because the advice is presented without information about the confidence of the advice. The purpose of this work was to develop a method to validate the advice of a computer by estimating the error of an artificial neural network output. A total of 1249 ECGs, recorded with computerized electrocardiographs, on patients who had undergone diagnostic cardiac catheterization were studied. The material consisted of two groups, 414 patients with and 835 without anterior myocardial infarction. The material was randomly divided into three data sets. The first set was used to train an artificial neural network for the diagnosis of anterior infarction. The second data set was used to calculate the error of the network outputs. The last data set was used to test the network performance and to estimate the error of the network outputs. The performance of the neural network, measured as the area under the receiver operating characteristic (ROC) curve, was 0.887 (0.845-0.922). The 25% test ECGs with the lowest error estimates had an area under the ROC curve as high as 0.995 (0.982-1.000), i.e. almost all of these ECGs were correctly classified. Neural networks can therefore be trained to diagnose myocardial infarction and to signal when the advice is given with great confidence or when it should be considered more carefully. This method increases the possibility that artificial neural networks will be accepted as reliable decision support systems in clinical practice.

Intelligent computer reporting 'lack of experience': a confidence measure for decision support systems.

The purpose of this study was to explore the feasibility of developing artificial neural networks that are able to provide confidence measures for their diagnostic advice. Computer-aided decision making can improve physician performance, but many physicians hesitate to use these 'black boxes'. If we are to rely upon decision support systems for such tasks as medical diagnosis it is essential that the computers indicate when the advice given is based on experience, i.e. give a confidence measure. An artificial neural network was trained to diagnose healed anterior myocardial infarction and to indicate 'lack of experience' when test electrocardiograms were different from the electrocardiograms of the training set. A database of 1249 electrocardiograms from patients who had undergone cardiac catheterization was used to train and test the neural network. Thereafter, the ability of the network to indicate 'lack of experience' was assessed using 100 left bundle branch block electrocardiograms, an electrocardiographic pattern that was excluded from the training set. The network indicated that 83% of the left bundle branch block electrocardiograms and 1% of the test electrocardiograms from catheterized patients were different from the electrocardiograms of the training set. All but one of the left bundle branch block electrocardiograms would otherwise be falsely classified as anterior myocardial infarction by the network. Artificial neural networks can be trained to indicate 'lack of experience', and this ability increases the possibility for neural networks to be accepted as reliable decision support systems in clinical practice.

Fuzzy sets in life sciences

Beginning with the first attempts to develop computerized systems in medicine and biology, it was recognized that one of the significant obstacles was the inherent uncertainty which arises as a natural occurrence of these types of applications. Unlike physical science applications, the life sciences are not amenable to straightforward algorithmic solutions for a number of reasons, including lack of complete understanding of mechanisms in biological organisms, inability to obtain complete information regarding the state of the organism, lack of precise ranges of normal values for physiological parameters, and complications caused by the interaction of several physiological systems functioning simultaneously. The most complex of these organisms, the human, is the most difficult to model. In fact, most efforts have been toward developing medical decision-support systems. Because of these inherent uncertainties, creative approaches were sought to develop both modeling and decision making systems in the life sciences. Initially, statistical and pattern recognition approaches were tried, followed by a decade of research into knowledge-based systems. Neural network modeling was also used beginning in the late 1960s. This approach met with little success but reemerged in the 1980s, with new algorithms and faster computers, to become one of the primary approaches in medical decision making. Although the early systems recognized the presence of uncertainty, it was usually dealt within a ad hoc fashion. Only in the last fifteen years have techniques from fuzzy logic been applied extensively in medical systems. However, in that short time, these systems have made significant inroads in modeling complex problems and producing practical decision support systems.