Fuzzy Cognitive Maps

Interfusion Services uses Fuzzy Cognitive Maps (FCMs) as an intelligent modelling tool for supporting planners, decision makers and managers in order to simulate what-if scenarios and help them assess and evaluate the potential impacts of various policy decisions. Fuzzy Cognitive Maps are a prediction tool that aim to capture and organize the knowledge of experts in complex social, economic, political and environmental problems in the form of a ‘mental model’, which reflects the complex domains of a given problem, as well as their internal relationships in an intuitive and dynamic manner.

Examples of their application include monitoring environmental effects, determining community facilities, and planning for urban development. In addition, they have been adopted to solve problems in scientific areas such as engineering and control, pattern recognition, energy, business and management, medical decision support, data prediction and forecasting, education, and software engineering.

Fuzzy Cognitive Maps combine the learning and classification abilities of artificial neural networks with fuzzy logic in order to visually depict expert knowledge with an acyclic directed graph composed of fuzzy nodes and edges. The graph’s nodes represent the concepts involved in the given problem (possible variables, states, events or actions), which can have either a positive or negative presence in the problem, and can be expressed quantitatively or qualitatively. The edges represent the causal relationships between these concepts either as a positive (excitatory) effect or a negative (inhibitory) effect. Once a problem is modelled as a FCM, the map is initialized and left to interact. The outcome provides insights of how the concepts of the problem are expected to change due to their interactions.

Policy makers can utilize this tool in order to examine how different measures can affect a problem, that is whether the measure achieves the desired outcome or whether the measure fails in adequately solving the problem.

We have experience in using Fuzzy Cognitive Maps in a variety of application domains, in particular:

– Predicting outcome of political referendums (case study: the Cyprus Problem) [1]

– Evaluating policy changes in order to improve tourism sustainability (case study: Municipality of Pegeia, Cyprus)

– Assessing designs of green spaces for the integration of young autistic children with non-autistic children (case study: City of Zagreb, Croatia)

– Organizing recreational activities (case study: City of Skopje, FYROM)

The latter three were developed during the FP7-ICT-funded project ‘FUPOL-Future Policy Modelling’. The Fuzzy Cognitive Maps where constructed with the help of domain experts from the three pilot cities and through their use allowed the responsible local authorities to evaluate different scenarios during the policy decision-making life cycle [2,3]. Moreover, the FCM for tourism sustainability served as a successful diagnostic tool for the Municipality of Pegeia as it allowed the town to pinpoint specific problem areas affecting tourists. For instance, the town was able to put in place smart environmental policy measures to reduce the number of cigarette butts on the beach (considered as the top litter problem for beaches). One of these measures was the installation of environmentally-friendly ashtray cone stands for tourists to use during their visit.

Furthermore, under the ERA-NET Urban Europe project ‘SmartGov’, we are currently undertaking the modelling for improving waste vehicle management for the Municipality of Limassol in Cyprus, and for routing the chaperoning of schoolchildren for the Municipality of Quart de Poblet in Spain.

References

  1. H. Mateou, C. Stylianou and A. S. Andreou, “Hybrid fuzzy cognitive map modeller: a novel software tool for decision making”, in Proceedings of the 4th IEEE International Workshop on Soft Computing as Transdisciplinary Science and Technology, pp. 851–862, IEEE Computer Society, May 2005.
  2. Neophytou and C. Stylianou, Fuzzy cognitive maps for urban policy design in developing countries, ch. 3a, pp. 124-131. E-governance and Urban Policy Design in Developing Countries. Nairobi, Kenya: UN-Habitat, 2015.
  3. Andreou, H. Neophytou and C. Stylianou, A review of the application of fuzzy cognitive maps in the policy decision-making life-cycle, ch. 8, pp. 129-148. Handbook of Research on Advanced ICT Integration for Governance and Policy Modelling. Hershey, PA, USA: IGI Global, 2014.