Chapter 1 Uncertainty in Dynamically Changing Input Data

The main objective of multiple criteria decision making models is to select an alternative, from a finite number, regarding a set of pre-defined criteria. Usually, this type of problems includes two main tasks, rating the alternatives regarding each criterion and then ranking them. Once a decision is made (alternative selected) the problem is solved. However, for situations involving reaching consensus or requiring several steps before reaching a final decision, we must consider a dynamic and adaptable decision model, which considers previous solutions. In this work we introduce multiple criteria dynamic decision making models (MCDDM) and discuss contributions to deal with the difficult problem of imprecision in dynamically changing input data. To illustrate the approach, a simplified example of au-tonomous spacecraft landing is presented. 1.1. Problem Statement In general, the aim of Multiple Criteria or Multiple Attribute Decision Making problems 1 (henceforth called multiple criteria) is to find the best compromise solution from all fea-sible alternatives assessed by pre-defined criteria (or attributes). This type of problems is widespread in real life situations1,2. A Multiple Criteria Decision Making (MCDM) prob-lem is usually modeled as a decision matrix, as depicted in Eq. (1.1). Basically, a MCDM model includes two phases3: (a) classify the alternatives regarding each criterion and then aggregate the respective classifications to obtain ratings per alternative; (b) order the alter-natives (ranking), where the highest rate usually corresponds to the “best ” alternative to be selected