Fuzzy logic has been serving the industry for decades by resolving the ambiguities that appear as a result of imprecise environment. High-stake decision-making processes require inputs from various stakeholders to incorporate. If the risk is high, as in the case of high investment decision making, a robust system of incorporating opinions from multiple stakeholders must be set in place in order to avoid any inconsistency or bad decisions. Fuzzy matrices and arithmetic can play a rescuer in such situations. In this chapter, the authors demonstrate a decision-making framework incorporating the use of fuzzy numbers and arithmetic to make critical decisions in strategic marketing and new product development. Forecasting in the domains of new products is an utmost complex and critical process because no relevant history is available owing to the product's ‘one-of-its-kind' nature. In such cases, computation via analogy is an interesting paradigm, which is also discussed in the chapter.
Top1. Introduction
Human has always been trying to model the behaviour of future events scientifically. A scientific approach to forecasting is based on the supporting facts and reasoning to come to a conclusion. However, the mantra behind every scientific approach remains the same and that is to reduce the uncertainty as much as possible. There exist a number of forecasting methods to deal with uncertainty in information. Usually, the choice of a forecasting method is determined by nature of the domain in which forecast has to be made and availability of the prior information or evidence. For example, there are different preferable methods for technological forecasting, financial forecasting and product sales forecasting. Also, the situations—when the evidences or the historical data is sufficiently available; when the historical data is absolutely unavailable, and; when the historical data is available for short periods only—influence the choice of a forecasting method.
Suppose an organization is ready with a design or a prototype of the innovation that it has produced, and huge amount of money and its image is at stake. Now, the organization decides that “it would be much beneficial, in implementing a promotional plan and deciding the production capacity of the upcoming setup, if an estimate, although imprecise, of the initial demand pattern is available.” This is that daunting forecasting problem where no prior data or historical observations are available to perform any analysis. It is not needed to mention that time series or regression analysis cannot be performed as no historical data is available. A special class of methods called qualitative methods is applicable in this context. Qualitative methods make use of expert opinion in forecasting and are thus subject to a number of biases and imprecision. Some techniques that can be used in qualitative forecasting are Delphi method, forecasting by analogy, scenario forecasting, sales force composite, executive opinion, and customer intentions (Hyndman & Athanasopoulos, 2013).
The authors propose an analogy detection methodology where a number of seemingly analogous analogy products are considered. It is considered a good practice to diversify the investments, as it minimizes the risk. Analogically similar practice is followed in the methodology. Instead of counting on a single analogy for forecasting, a set of analogies is often preferred. A few possible reasons behind using multiple analogies may be:
- 1.
Decision makers often recall multiple resemblances, not one.
- 2.
It is riskier to put the entire stake on a single analogy.
- 3.
Different analogies may resemble in different aspects.
Because, every analogy cannot possess equal resemblance, there is a need to identify the analogies with significant resemblance. Identifying such analogies is not a trivial task. The process of selecting a set of good analogies, in fact, requires complex decision making ability at the end of decision maker (DM). Thus, a well structured approach for deciding the appropriate analogies is warranted at this juncture.
Top2. Supplementary Techniques For Forecasting In Absolute Absence Of Data
Forecasting is indispensable in an organization because crucial decisions are often required to be made that are based on the decision makers belief in forecasting results. A reverse version of the said statement is also valid sometimes i.e. decision making sometimes become essential for generating forecasts. Decisions are made to be taken in the presence of multiple, often conflicting, criteria; had there been a single criterion, no decision would have to be taken at all (Shi, Wang, Kou & Wallenius, 2011). Decision making is not a trivial process. It requires tremendous efforts from the side of a decision maker. Therefore, a structured way of decision making is being practiced and studied independently under the name of Multiple Criteria Decision Making (MCDM). A number of approaches and models exist in MCDM literature that make the decision making structured and the outcomes more reasonable. MCDM can broadly be classified into two categories: Multiple Attribute Decision Making (MADM) and Multiple Objective Decision Making (MODM). MADM deals with rating of the alternatives to form a rank, based on preferences (Kenevissi, 2014). MODM, on the other hand, provides a mathematical framework for designing a set of decision alternatives.