Decision and Cognitive Sciences Research Centre

Research themes

Research themes

Whilst it is neither possible nor desirable to be exclusive about specific research themes, ten initial research themes have been identified, each with a dedicated theme leader(s). The research themes will form distinctive yet inter-linked and synergistic focus points for organising, promoting and conducting research to achieve the DCS objectives. The theme leaders will be proactive and take leadership in developing and coordinating research in their theme areas. They will be a first point of contact for research collaboration and project development in their areas of interest. The ten research themes and theme leaders are listed in the following table.

Theme leader: Jian-Bo Yang

This theme has three key research focuses: evidential reasoning decision analysis, belief rule based risk analysis, and performance assessment and planning. The first focus will build on our current strengths of research on multiple criteria decision analysis with both quantitative and qualitative information under uncertainty, which has been extensively supported by research councils and industry. This research will be focused on investigating the interrelationships between statistical decision-making and evidence-based decision-making. The ultimate goal is to develop an evidential reasoning decision analysis methodology and theory to support informed and robust decision making in a wide range of areas. The second research focus will be based on our current research projects funded by EPSRC and other bodies to provide generic frameworks, models, methods and tools to facilitate risk, safety and security analysis and decision making. In many social and engineering systems, it is difficult to estimate the likelihood of occurrence of a risk event, its direct impact and potential vulnerability. This is largely due to the need to analyse many factors that are often associated with various types of uncertainty such as subjectivity, incompleteness and even lack of understanding and data. The frameworks, models, methods and tools will be developed to help handle such uncertainties and applied to various areas such as supply chain risk and security assessment, dynamic clinical risk assessment, project risk assessment, financial risk analysis, corporate risk analysis, and early warning systems for social crises.

The third research focus is to investigate models, methods and processes to integrate performance assessment and performance planning. While performance measurement is becoming increasingly more common in many organisations such as NHS hospitals, it has not yet become performance management. This research will be focused on applying analytical methodologies like data envelopment analysis and multiple objective linear programming to help reconcile management control and management planning in a consistent way with the decision makers’ preferences taken into account in an interactive fashion.

Theme leader: Andrew Howes

We have an established research strength in computational and mathematical models of human behaviour. Our focus is on people as adaptive decision makers: People choose actions, gather information and plan, under psychological constraints and in response to perceived utility. We apply computational models of decision making to real world problems so as to predict individual performance. This work has recently been sponsored by NASA and by the Office of Naval Research (ONR) and is conducted in collaboration with the University of Michigan.

Theme leader: John Keane

The generation and capture of very large complex datasets (structured, semi-structured, unstructured) has led to commensurate need for more sophisticated analysis methods. Data mining is the process of extracting patterns from data. Data mining is becoming an increasingly important tool to transform this data into information. In turn, text mining refers to the process of deriving high-quality information from textual data.

Both data and text analytics algorithms search for unknown information within such datasets. Specific work involves hybrid approaches, itemset mining, classification algorithms, fuzzy systems, rough sets and language modelling Bayesian DSS. A number of domain areas are of interest, including health informatics, statistical disclosure, price modelling, electricity forecasting, security applications and customer relationship management.

Theme leader: Dong-Ling Xu & Weigang Wang

Built on our experience and strengths in the development of interactive web based applications, this research theme is aimed at developing a live web platform to showcase some of the most recent advancements in decision and cognitive sciences. The main activities are to explore the potential of applying the most advanced Internet technology to implement the theory, methodology, framework, techniques, ideas etc. developed by the members of the centre and beyond to support organisational and societal decision making. This includes

  • the development of web based problem structuring and modelling tools,
  • the development of web based decision support systems such as evidence based web voting, opinion survey, impact assessment, performance assessment, and risk assessment,
  • investigating how to facilitate wide participation and shared understanding of group decision making activities, and
  • investigating distributed facilitation for web-based group decision making processes.

Theme leader: Simon French

Here in the UK and, indeed, across the world, there is a widespread push to involve the public in societal decision making. And it is not just in relation to national, regional and local government. Regulators and agencies at arm’s length from government are running more and more stakeholder workshops, citizen’s juries, etc. This is particularly the case in relation for societal risk management. However, while these developments have been explored and justified from many perspectives – political, philosophical, social, psychological, managerial, and so on – there has been less attention to the decision analytic perspective. How does one design a participation process to involve citizens and stakeholders truly in a societal decision and how does one evaluate its success?

This theme within DCS brings together a number of strands of activity in which decision analysis has a lot to contribute to effective participation, addressing research questions such as:
  • how to involve most effectively a range of stakeholders and citizens in formulating issues and framing the decision;
  • how to present and explore analyses with a range of stakeholders and citizens, probably varying in, e.g., cognitive abilities and cultural perceptions;
  • how to vote on, identify a consensus or negotiate a solution which is acceptable in some sense, mindful of all the paradoxes in the domain of democracy and collaborative decision making;
  • how to design the most cost-effective overall participation process in the broadest sense of cost effectiveness.

Theme leader: Ludmil Mikhailov & Dong-Ling Xu

The area of Intelligent Decision Support Systems (IDSS) is interdisciplinary in nature, bridging Artificial Intelligence, Decision Science and Information Systems. Our research on IDSS focuses mainly on theoretical methods for multiple-criteria decision-making and application of intelligent technologies, such as expert systems, fuzzy logic systems and genetic algorithms. We are particularly interested in intelligent support to decision-making under uncertainty, based on the Dempster-Shafer evidence theory and fuzzy logic, and multiple objective optimisation by evolutionary computing. An important objective of our research is the application of those methods for the development of software tools for decision support.

The main directions in our research on IDSS are:
  • Belief rule based expert system, based on a novel approach on evidence reasoning. The developed IDSS is used for solving many problems, such as clinical decision support and engineering system fault diagnosis. The system is used by practitioners, decision analysts and researchers from over 30 countries including organisations such as General Motors Company, Belgian Nuclear Research Centre, and Hong Kong Productivity Council.
  • Fuzzy logic methods for reasoning under uncertainty and their application to the development of IDSS. Prototypes of fuzzy IDSS have been applied to solving decision-making problems, such as evaluation of services, partnership selection in virtual enterprises, outsourcing of IT services, risk analysis and evaluation of offers in telecommunication markets.
  • Multiple objective optimisation by evolutionary computing. This approach to multicriteria decision-making allows obtaining a set of Pareto-optimal criteria weights from inconsistent expert judgements, by multiobjective evolutionary algorithms. Our research in this area aims to develop a new class of interactive IDSS, which could be applied to group decision making.
  • Evolving IDSS. Those systems utilise learning and self-organization principles, in order to adapt in a dynamic environment. We are particularly interested in fuzzy evolving IDSS, which evolve their fuzzy rules by learning from data. Our prototype of evolving IDSS for intrusion detection successfully detects and classifies various types of security attacks.
  • Applying intelligent software tools to support complex decision analysis. Some recent applications include nuclear waste management, environmental management, risk assessment, and organisational self-assessment.

Theme leader: Paul Jackson

This theme will examine decision making processes in situations where information is highly complex (many aspects which are interdependent), highly uncertain (fuzziness and unpredictability), and also dynamically updated. We will concentrate particularly on decision processes within teams, which may be top management teams or project teams. We will examine the consequences of differences in the cognitive models of team members (some scholars argue that diversity is a benefit, while others describe the problems which can arise from different cognitive models), the ways in which new knowledge can arise out of people working together in teams, the emergence of shared cognitive models within work teams and the consequences of this for team effectiveness, and ways in which decision support tools can facilitate effective decision making.

Theme leader: Oscar de Bruijn & Andrew Howes

Information Technologies provide opportunities for reshaping the work environment. Social Media provide tools with which people can maintain connectivity and, potentially, reap rewards associated with enhanced social capital and empowerment. Our approach to understanding the impact of these technologies focuses on the individual choosing how to maintain social and work relationships through technology. For example, we are investigating the role of reciprocal action and reputation in the maintenance of online communities and the way in which technologies might more directly support these key social functions. A crucial role in this is played by the ability of people to make sense of information and situations. This puts sensemaking research into the centre of what we are doing. In this respect we focus on the role of information visualisation and online personas in guiding decision making behaviour as key research areas for the theme.

Theme leader: Ser-Huang Poon

This theme has two key focuses: optimisation and simulations both underpinning important decisions in the finance industry. There are many financial decisions such as wealth-bench portfolio management and retirement pension plan that involve long horizon, multi-period non-linear constrained optimisation. The classical gradient search routines failed as corner solutions are abundant. The genetic algorithm search is too time consuming and not suited for practical application in the industry. Recently, we have some success in applying the “Belief Rule Based” algorithm in a 9-asset, 60-year portfolio optimisation that can benchmark against industry standard. The plan is to capitalise on this early success and apply BRB in more advanced problems.

The second research focus is to find practical solution to speed up risk measure calculation that involves simulations. The valuation of many financial structural products relies on simulation because of the number of stochastic variables involved and the complexity of the structure imposed. This makes the risk management of these structural products almost an impossible task, as stress testing (a typical regulatory required procedure in e.g. Basel) also involves simulations of scenarios. This theme will tackle the project from two angles. The first is to find mathematical approximation to by-pass the first-stage simulation. The second step is to find efficient schemes (from a statistical perspective or from a computer science perspective) to obtain the risk measures that are associated with these structural products.

Both focuses are affiliated to the Alliance MBS Marie Curie Framework 7 Early Stage Researcher Training Grant with industry partners as collaborators.

Theme leader: Jian-Bo Yang, David Bamford and Andrew McMeekin

With the growing world population and increasingly scarce natural resources, it is paramount for organisations to measure and assess performance systematically and improve productivity, sustainability and innovation continually. In recent years, there has been increasing pressure for organisations to institutionalise measurement, assessment and improvement of performance, sustainability and innovation due to the convergence of two forces: (1) increased demand for accountability on the part of governing bodies, the media, and the public in general, and (2) a mounting commitment of managers and government agencies to focus on results and work more deliberately to strengthen performance, sustainability and innovation. This research theme is built on the DCS’s strengths in the theoretical and methodological research in these areas, which has been supported by EPSRC, EC and industry in the past, to conduct applied research focused on the following Alliance MBS priority areas.

  • Performance assessment and improvement – Organisational performance assessment needs to take into account a large number of performance indicators, is associated with the identification and collection of domain specific knowledge, best practices and industry standards, depends on the measurement of different performance indicators, and requires systematic approaches for rationally aggregating performance indicators for performance-related decision making. This research will be focused on applying theories, methodologies and tools developed at DCS to support practical performance assessment and improvement through engagement with businesses via KTP and consultancy projects.
  • Evaluation of sustainability – Sustainability can be measured and evaluated at both micro and macro levels. At micro level, sustainability programmes initialised by an organisation needs to be evaluated by comparing created opportunities with potential risks, such as for the product carbon labelling programme that we have investigated for Tesco. On the other hand, sustainability can be regarded as an important dimension of an organisation’s performance indicator framework. At macro level, sustainability can be related to strategic issues like sustainable economic development, energy policy, climate change, etc. A common challenge in sustainability evaluation is how to aggregate a large number of sustainability indicators to support strategic decision making. This research will be focused on the further development of rigorous yet pragmatic methods and tools to facilitate sustainability evaluation related to strategic issues in the real world.
  • Evaluation of innovation – Evaluation in the field of innovation is increasingly necessary. It is a complex process and needs to be promoted. The previous research in this area at DCS was supported by EC and focused on helping SMEs to evaluate their innovation capabilities by developing frameworks and tools for systematically evaluating innovation level, strategy, process, culture, etc. On the other hand, evaluation of innovation programmes is important to support innovation policy making for government departments and other agencies. Our future research will also be directed to developing systemic approaches and frameworks to support evaluation-based innovation policy development.

The DCS management team consists of the Directors and Coordinators who are in charge of the day to day running of DCS. The management team and the theme leaders form the DCS management committee, which meets to set the policies and directions for DCS. The theme leaders are responsible for organising project meetings to initiate new research projects and manage existing research projects as appropriate.