A Review of the Application of Fuzzy Cognitive Maps in the Policy Decision-Making Life Cycle

Introduction

Modern-day policy decision-making has become more complicated and involves greater risks compared to past decades. This is largely down to the increase in the number of variables that need to be taken into account when making decisions concerning the adoption of a new policy or modifying the content of an existing policy. For instance, recent social issues and environmental laws have influenced both the subject matter of policies as well as the process of policy decision-making. Additionally, the availability of more experts, the increase in the accessibility of data, as well as the establishment of social media as part of our everyday lives has required the policy decision-making process to be altered to accommodate these new aspects. Furthermore, citizens are now given greater opportunities to have their opinions heard, as well as means to marshal support or rally opposition towards various policy decisions, especially with the increasing use of Internet and mobile/smartphone applications and social media sites.

Part of the agenda setting stage of the policy-making life-cycle (discussed in detail in Chapter 5) involves defining the problem to be addressed and establishing the need for a new policy or a change in an existing policy. Once these activities are completed, the analysis stage begins – challenges and opportunities that are associated with an agenda item are defined in detail and recorded in a draft policy document. In order to specify exactly what a policy is supposed to accomplish, it is important to specify its goals and objectives, including the ways to measure its impact on economic, socio-cultural, environmental or any other relevant concerns defined in specific criteria and indicators. Additionally, it is equally important to gather as much knowledge and expertise as possible from different sources, such as research-based or scientific knowledge from previous projects and policy implementations, as well as statistical information, to help in upcoming simulations and forecasts performed later on in the analysis stage. After these activities have been carried out, policy-makers will be left with a number of approaches or strategies from which they will need to decide which one will make the better policy to implement in the next stage of the life-cycle. To evaluate all these options, policy-makers have to model and simulate the impact of each policy and match the results against the criteria and indicators defined previously. Analysing the characteristics of the options will then help determine the final strategy to implement as a policy.

Constant changes in transportation, economics, family life, medical practices, and many other aspects can heavily influence the way a policy is perceived by citizens in terms of quality factors. Added to the fact that the process of drafting and implementing policies is inherently very challenging and highly complex, this has left policy decision-makers struggling to successfully exploit the knowledge and expertise of specialists, and to adequately predict the impact of a policy and how citizens will perceive it. This is also the main reason why very few specific modeling and simulation methods have been developed and adopted to help with policy decision-making in the past. Recently, however, policy decision-makers have shifted towards using intelligent techniques to help them overcome these challenges. One of the most popular techniques used is fuzzy cognitive maps – an expert-based qualitative modeling technique. Generally speaking, fuzzy cognitive maps (FCMs) are used to depict a mental representation of the factors existing within a given problem. These factors can be states, notions, events or any concepts that possess certain characteristics and qualities. Causality is then incorporated by connecting these factors together based on the influence each factor exerts on another, resulting in a directed graph of concepts and relationships known as a cognitive map. Once the map is formed, fuzziness is introduced in the connections so that causality (i.e., each edge/relationship) is denoted as matter of degree to signify the strength of the influence between two factors. To help decide the factors to be included, the causality relationships and the degree of strength of each influence, domain experts are consulted providing their knowledge and expertise. As a modeling technique, FCMs can help assess the current situation of a problem, but their main advantage lies with their ability to predict the future state of a problem by simulating the interactions of the factors in the form of a fuzzy feedback model of causality. Hence, FCMs are an ideal inference tool to perform what-if scenario analysis (Glykas, 2010).