Democritus University of Thrace (DUTH)
Development and application of hybrid fuzzy-statistical and artificial intelligence methods in the design, operation and management of transportation systems and infrastructures
Abstract
dc:descriptionFuzzy logic and probability theory are both useful for dealing with uncertainty, various models have been built using these two paradigms. However, there has been difficulty in combining the benefits of these two types of models because of the difference in approach and model formulation. For example, various fuzzy versions of probabilistic models have been built but their use has been limited. The reason is that experts in probabilistic models may not have the required knowledge in fuzzy logic to understand and use such fuzzy models. Similarly, different ways have been suggested to include probability and statistics in fuzzy logic models, but such models have had very limited applications. This is also because experts in fuzzy logic may not possess the adequate statistical background to understand these types of models. The goal of this thesis was to solve this problem of methodological difference between statistical and fuzzy models.In contrast to most works which normally build new fuzzy or statistical versions of original models, this thesis proposes to build hybrid fuzzy-statistical models that use some important elements of one model in the other without significant change in the structure of the original models. The advantage of the proposed approach is that experts in any of the models (fuzzy or statistical) can continuing using such models without the need to acquire significant new knowledge in the other domain.It is shown that the proposed approach gives comparable or better results to the original models. In the various applications, the proposed models proved to be highly flexible since already existing models like linear regression, neural networks and fuzzy logic models can be easily transformed to fuzzy-statistical models. The need for additional knowledge is minimised and the proposed models can be used by anyone already familiar with linear regression, fuzzy logic or neural network models. Also, it shown that the proposed models can handle uncertainties related to randomness and fuzziness and are particularly suited to handling most transportation decision making problems which are usually made under high uncertainty.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Adjenughwure, Kingsley
Subjects
dc:subject × 24- Ασαφής λογική
- Στατιστική
- Διαστήματα εμπιστοσύνης
- Συγκοινωνιακός σχεδιασμός
- Διαστήματα πρόβλεψης
- Τεχνική νοημοσύνης
- Fuzzy logic
- Statistics
- Prediction intervals
- Artificial intelligence
- Transportation panning and economics
- Transportation statistics
- Fuzzy statistics
- Fuzzy inference systems
- Φυσικές Επιστήμες
- Μαθηματικά
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Πολιτικού Μηχανικού
- Natural Sciences
- Mathematics
- Computer and Information Sciences
- Engineering and Technology
- Civil Engineering
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/45924