National Technical University of Athens (NTUA)
Υπολογιστική προσομοίωση αστικής επέκτασης με χρήση ασαφών κυψελοειδών αυτομάτων
Abstract
dc:descriptionIt is of major importance that urban growth occurs in a planned way maximizing the benefits for urban population while minimizing both environmental and economical cost. This requires accurate and realistic estimations of the urbanization process and this is what urban models do. The term modeling refers to creating a strictly defined analog of real world by subtraction and provides simulations and future projections under identifiable assumptions to suggest what the future might be like. Nevertheless, there is no rigorous framework for modeling such a spatio-temporal phenomenon as urban growth since there lies great inherent spatial, temporal and decision-making heterogeneity, which results from socio- economic and historical heterogeneity itself. Apparently, our knowledge, is not really describing urban growth dynamics in general, but instead the part of the urban growth dynamics that have already occurred and have been observed and experienced. What is more, knowledge about the operational scale(s) of urban form and process, and the interaction and parallelism among different scales, is poor, partially due to the recurring problem of lacking spatially detailed data. Apart from dealing with the above mentioned issues, for a model to be useful it should not only provide accurate estimations but also express both its results and its mechanism – relations, interactions and assumptions - in an open, visible, explicit and comprehensible way in order to be challenged by knowledgeable people. For these reasons a modeling framework has been developed, CaFe (Cellular Automata - Fuzzy Engine) whose goal is to: • Be reliable and realistic. This means to provide accurate estimations for the future urban cover and reproduce efficiently the spatio-temporal patterns of the urban growth dynamics • Be open and adaptable to the user’s requirements. This is to: - describe the model’s mechanisms and the workflow in simple and comprehensible terms and - support the population of alternative scenarios • Be easy to use and resilient to the case study limitations, more specifically to: - sustain a generic versatile form, disengaged from severe data limitations and - use knowledge base in such form that it can be reduced or extended. The proposed model is rule-based and supports spatio-temporal rules that may be either data-driven or empirical. As a result the knowledge base may fit better to reality allowing the user to overcome possible data limitations - which lead to lacking of specific knowledge - by using exogenous knowledge adapted to the model according to empirical similarity patterns. What allows the desired objectives to be accomplished is the combination of the descriptive strength of Fuzzy Logic and the computational strength of Cellular Automata. The model applies an innovative workflow to connect the input and the output variables. All input variables are merged in a single thematic layer, the ‘urbanization suitability’, which defines the potentials of each area to develop. To do so, the model utilizes Sensitive Sum; a new fuzzy operator that is developed to employ a dynamic parallel connection between the effects of separate input variables while taking into account their statistical correlation. As a result the model does not require certain variables/data to run while it implements a reducible/extensible form of Knowledge Base which can include both data-driven and empirical rules. The simulation engine is separated by the suitability calculator and incorporates advanced Cellular Automata techniques. These techniques apply spatio-temporal multi radius transition functions and support action in distance. Specifically, on top of the ‘traditional’ cellular automata, CaFe introduces directional cellular automata and adopts pseudo-agent behavior. CaFe is applied in the broader area of Mesogia in east Attica (Athens-Greece) using real data for 1988, 2000 και 2007. More specifically, the model was structured and calibrated for 1988-2000 and was applied for 2000-2007 και 1988-2007. During these periods, urban cover grew by 66%, 66% and 200% respectively, composing thus a challenging case study. Nevertheless, CaFe manages to map urban cover successfully and to efficiently simulate urban growth while scoring high fitting indicators and retaining a stable average error.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2011
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Mantelas, Lefteris
- Μαντέλας, Ελευθέριος
Subjects
dc:subject × 16- Αστική επέκταση
- Κυψελοειδή αυτόματα
- Κυψελοειδή αυτόματα κατά κατεύθυνση
- Χωρο-χρονική συνάρτηση μετάβασης με μεταβλητή ακτίνα
- Λογική της ασάφειας και ασαφή συστήματα
- Ασαφής τελεστής ευαίσθητο άθροισμα
- Urban growth
- Cellular automata
- Directional cellular automata
- Spatio-temporal multi radius transition function
- Fuzzy logic and fuzzy system
- Sensitive sum fuzzy operator
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Πολιτικού Μηχανικού
- Engineering and Technology
- Civil Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/24446