Name: RODOLFO LOURENZUTTI TORRES DE OLIVEIRA
Publication date: 17/06/2016
Advisor:
Name |
Role |
|---|---|
| RENATO ANTÔNIO KROHLING | Advisor * |
Examining board:
| Name |
Role |
|---|---|
| RENATO ANTÔNIO KROHLING | Advisor * |
| CLAUDINE SANTOS BADUE | Internal Examiner * |
Summary: Multi-Criteria decision-making (MCDM) problems are becoming increasingly complicated over the years. Consequently, there is a growing necessity for more resourceful methods to support decision making in settings that are more complex. The methods must now deal simultaneously with several different aspects of MCDM problems. One of the most common aspect is the group decision making. When a team of decision makers is faced with a problem, it is natural to assume that due to diversity of their backgrounds and experiences they may require different ways of providing their evaluations and opinions. However, many MCDM methods make prohibitive, and sometimes unreasonable, assumptions. For example, most of the methods for group MCDM assume that all the decision makers agreed on a unique criteria set to evaluate given alternatives. Another issue is the necessity to process heterogeneous data types. Different situations may require different data representations. For instance, for precise information one may want to use crisp numbers, while for stochastic uncertainty random variables may be more appropriated. The necessity of processing heterogeneous data becomes even more evident when dealing with groups of decision makers, due to the different experiences and backgrounds. Additionally, more often than one may think, there is the problem of criteria interaction. Failing to consider the interaction between the criteria can lead to an overweight of certain traits of the problem resulting in an unbalanced analysis. A known approach in literature to deal with criteria interaction is the Choquet integral. However, the Choquet integral brings a different and very complicated problem with it, which is the fuzzy measure identification. The number of coefficients in a fuzzy measure grows exponentially with the number of criteria. On top of all that, the alternatives may be affected by underlying factors that are not being directly considered. The change of these factors may cause changes in the context that the problem is being evaluated. These factors can be deterministic, such as time or place, or stochastic, such as the weather. Assuming a static environment can lead to poor modelling of a problem. Aiming to address all these problems at once, this thesis start presenting variant of the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) and TODIM (an acronym in Portuguese for Interactive and Multi-Criteria Decision Making) methods, named Intuitionistic Fuzzy Random TOPSIS (IF-RTOPSIS) and Intuitionistic Fuzzy Random TODIM (IF-RTODIM), which are able to process intuitionistic fuzzy data in dynamic environment. The strategy used by the IF-RTOPSIS and IF-RTODIM is then used by more general methods, named Group
