About Scenario Planning and Cross-Impact Analysis methods

Raising a large company which competes in our rapidly changing market is a hard task, specially nowadays due to general financial crisis which makes potential customers to be very careful regarding large investments and important dealings. It is thus of crucial importance for enterprises to be prepared for changes in order to anticipate the future and better adapt to new business environments and demands.

Scenario Planning [1], [2], [3], [4] has been a very widely employed technique for this purpose from long time ago, as shown in [5], [6] and more recently, in [7], [8]. It can be defined in a broad sense as a family of tools that emphasize and stimulate creative thinking of company managers, helping them imagine a number of feasible scenarios as systematically as possible. According to [9], a scenario is an internally consistent view of what the future might turn out to be - not a forecast, but a possible future outcome. A very similar definition is given in [10], emphasizing the fact that perspectives reflected by scenarios can serve as a basis to make decisions in the present. In order to get such internally consistent descriptions, further techniques are required to provide a clear understanding of the systems we want to analyze, which are typically complex and constituted by many interrelated components or factors.

Obtaining such a description is difficult because most of the times, there is no mathematical model able to capture the whole system. Sometimes it is possible to have a good model of an isolated part only, but that is not the original goal of scenario analysis. Therefore, a way to grasp a system that is too complex to be intuitively understood, and for which there is no well-founded mathematical model, is required. Among the variety of Scenario Planning techniques used to accomplish this, we can find Interactive Future Simulation, Interactive Cross Impact Simulation, Trend Impact Analysis (TIA), Cross Impact Analysis (CIA) and Fuzzy Cognitive Maps (FCM) [8]. TIA and CIA share the fact that they both employ probabilities provided by experts, but in TIA they represent probability of deviation from a model obtained from historical data, while in CIA [11] they are prior probabilities of the events involved in a scenario. Finally, the application of FCM for developing scenarios was done for the first time in [7]. FCMs capture causal relations in a weighted directed graph, which allows for the study of indirect relations and loops.

CIA methods in particular can help a lot in uncovering the role of the components within the system. This information can be very useful for a company, as it gives the managers an accurate view of which components they should modify in order to achieve the desired outcomes from other parts of the system that are affected by the formers. Such interrelated inner constituents are traditionally called factors in this context.

To achieve this, CIA methods normally take the relationship between each pair of impact factors as input. Such relations are provided by experts in the form of probabilities, either conditional or marginal. By applying suitable mathematical procedures to the input, these methods eventually find which scenarios are the most probable. The operations carried out are such that not only direct relationships between pairs of factors are considered, but also indirect relationships through one or more intermediate factors.

We will focus on a variant of CIA known as MICMAC (Matrice d’Impacts Crois´es Multiplication Apliqu´ee a une Classement [12]), which was employed in the last decade for Structural Analysis. This kind of analysis is aimed at identifying the key variables of the system under study [13]. The input consists of a non-probabilistic measure of the strength of the relation between every two variables considered in the system, according to a panel of experts. The output is not the probability of occurrence of scenarios, but a ranking indicating the relative importance of the variables considered at the input. We will further elaborate on this method in the next section. As stated in [14], MICMAC presents some drawbacks. Firstly, it does not cope with the vagueness inherent to experts’ mutual influence assessments, as it cannot be properly captured using crisp numbers. Secondly, the results are always numerical but lack a clear interpretation, beyond a simple scoring or ranking that are meaningless and little explanatory. Finally, the output consists of relative information, thus it is not possible to measure the relevance of a variable in absolute terms but only in relation to the rest. In addition to this, and despite the existence of many different variants of MICMAC presented in the literature, some of which are reviewed in the next section, almost none of them is publicly available, neither in a ready-to-use free implementation nor as a programming library.

Here we deal with a recently published Fuzzy Linguistic extension of MICMAC called FLMICMAC that solves the aforementioned shortcomings and more importantly, we present for the first time a ready-to-use implementation of our proposal in a web system and a Java programming framework. We demonstrate the value of our contribution by solving a real example using our web system.

References:
[1] P. W. van Notten, J. Rotmans, M. B. van Asselt, and D. S. Rothman, “An updated scenario typology,” Futures, vol. 35, no. 5, pp. 423–443, 2003.
[2] M. Godet, “The art of scenarios and strategic planning: Tools and pitfalls,” Technological Forecasting and Social Change, vol. 65, no. 1, pp. 3–22, 2000.
[3] G. Wright and P. Goodwin, “Decision making and planning under low levels of predictability: Enhancing the scenario method,” International Journal of Forecasting, vol. 25, no. 4, pp. 813–825, 2009.
[4] C. Varum and C. Melo, “Directions in scenario planning literature: A review of the past decades,” Futures, vol. 42, no. 4, pp. 355–369, 2010.
[5] R. E. Linneman and H. E. Klein, “The use of multiple scenarios by U.S. industrial companies: A comparison study, 1977 - 1981,” Long Range Planning, vol. 16, no. 6, pp. 94 – 101, 1983.
[6] P. Malaska, M. Malmivirta, T. Merist, and S. Hans´en, “Scenarios in Europe - Who uses them and why?” Long Range Planning, vol. 17, no. 5, pp. 45 – 49, 1984.
[7] A. Jetter and W. Schweinfort, “Building scenarios with Fuzzy Cognitive Maps: An exploratory study of solar energy,” Futures, vol. 43, no. 1, pp. 52 – 66, 2011.
[8] M. Amer, T. U. Daim, and A. Jetter, “A review of scenario planning,” Futures, vol. 46, no. 0, pp. 23 – 40, 2013.
[9] M. E. Porter, Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press, 1985.
[10] P. Van Notten, “Writing on the wall: Scenario development in times of discontinuity,” Ph.D. dissertation, University of Maastrich, Masstrich, The Netherlands, 2005.
[11] T. Gordon and H. Hayward, “Initial experiments with the cross impact matrix method of forecasting,” Futures, vol. 1, no. 2, pp. 100 – 116, 1968.
[12] J. C. Duperrin and M. Godet, “M´ethode de hierarchisation des elements d’un sisteme,” Rapport Economique du CEA-R-4541, 1973.
[13] J. Arcade, M. Godet, F. Meunier, and F. Roubelat, “Structural Analysis,” in Futures Research Methodology, version 3.0, J. C. Glenn and T. J. Gordon, Eds. The Millenium Project, 2009.
[14] P. J. Villacorta, A. D. Masegosa, D. Castellanos, and M. T. Lamata, “A new fuzzy linguistic approach to qualitative cross impact analysis,” Submitted to Applied Soft Computing, 2012, under review.