This chapter summarizes the main scientific results.
7.1 Thesis I: Domain-Specific Modeling and Model Processing
Related key publications: [Asztalos et al, 2007] [Lengyel et al, 2005] [Lengyel et al, 2006] [Lengyel et al, 2008]
[Lengyel et al, 2015c] [Lengyel and Charaf, 2015d] [Lengyel and Charaf, 2015e] [Levendovszky et al, 2005a]
[Levendovszky and Charaf, 2005b] [Mészáros et al, 2009] [Mezei et al, 2006] [Mezei et al, 2007].
The results of Thesis I are the followings:
Supporting Domain-Specific Modeling
‒ A methodology that supports to define the concrete syntax (appearance) of visual domain-specific languages. A domain-specific language and its application to define the concrete syntax (appearance) of visual domain-specific languages.
‒ The methodology applies model processing to combine the abstract syntax (defined by the metamodel) and the concrete syntax. The result is a visual domain-specific language that can be used for modeling.
Supporting Dynamic Behavior of Domain-Specific Languages
‒ A methodology that supports the dynamic behavior (animation) definition of domain-specific languages. The methodology contains the event-driven architecture that supports the definition of the dynamic behavior.
‒ A methodology that applies model transformations to execute the models describing the dynamic behavior with.
Model Transformation-Driven Software Model Processing
‒ A methodology that supports the graph rewriting-based model processing. Models are represented as graphs and the model processing is defined as a sequence of graph rewiring rules. It has been proven that the method provides efficient algorithms to find the left-hand side graphs in the processed models, and to replace these matches with the right-hand side graphs of the rewriting rules.
‒ The algorithms of the methodology facilitate the followings: (i) based on the metamodel multiplicities it is decidable whether valid instance models can be constructed or not, (ii) in the context of the inheritance relation, the identification of the topology relations between the metamodel and instance models. Conditions have been worked out (i) to decide the applicability of the rewriting rules and (ii) to analyze the serial and parallel independence of the rewriting rules and decide whether the rules can be executed in parallel and/or in different order. The algorithms and the conditions provides offline (before the transformation execution) analyzing possibilities of the rewriting rules.
Validating Model Transformations
‒ A methodology that supports the online validation of graph rewriting based model transformation. The validation is performed by the pre- and postconditions assigned to the rewriting rules. Based on the methodology, the execution steps of the rewriting
rules are as follows: (i) finding a topological match according to the left-hand side graph of the rewriting rule, (ii) validating the preconditions on the matched sub-graph, (ii) performing the rewriting (model processing), and (iv) validating the postconditions on the result of the rewriting.
‒ Algorithms with proven correctness, which facilitate the validation of both the rewriting rules and whole model transformations.
‒ A methodology to analyze the behavior of graph rewriting-based model transformations. The analysis is supported by a control flow language, which facilitates to define the exact execution order of the rewriting rules, and by the algorithms that make possible to contract transformation rules.
7.2 Thesis II: Increasing the Efficiency of Mobile Platforms
Related key publications: [Aczél and Charaf, 2005] [Braun et al, 2015] [Ekler et al, 2008a] [Ekler and Charaf, 2008b] [Ekler et al, 2010] [Ekler et al, 2015] [Fitzek et al, 2014] [Forstner et al, 2005] [Kelényi et al, 2007]
[Kundra et al, 2015] [Pahlevani et al, 2014] [Pándi and Charaf, 2013] [Pándi and Charaf, 2015] [Vingelmann et al, 2011].
The results of Thesis II are the followings:
Analyzing Mobile Peer-to-Peer Networks, Retrieving Semantic Information
‒ Analyzing unstructured peer-to-peer networks. A methodology, which supports that based on semantic information predict the probability whether the required information would be provided (the query could be answered).
‒ The method to design data structures, which support the methodology. Algorithms that based on semantic information build and utilize the data model.
Examining Mobile Platforms-related Social Networks
‒ A methodology that supports to model and discover the relations between social networks and mobile device phonebooks.
‒ Analyzing the scalability of mobile-based social networks. A methodology that supports the scalability of mobile-based social networks.
‒ A model that supports the estimation of energy consumption of mobile devices with limited resources.
Energy Efficient Mobile Peer-to-Peer Solutions
‒ A methodology facilitating that instead of partitioned data transfer to mobile devices the data is transferred in bursts (larger packages). The solution provides significant energy saving for mobile devices.
‒ Algorithms and a protocol that makes data transfer to mobile devices more energy efficient.
7.3 Thesis III: A Model-Driven Methodology for Supporting Multiple Mobile Platforms and IoT Devices
Related key publications: [Charaf, 2013a] [Charaf, 2013b] [Charaf et al, 2014] [Ekler et al, 2015] [Lengyel et al, 2015a] [Lengyel et al, 2015b] [Lengyel et al, 2015c] [Lengyel and Charaf, 2015d].
The results of Thesis III are the followings:
Modeling Common Aspects of Different Mobile Platforms
‒ Investigated and selected the technology that based on the research goals could be applied to work out the mobile applications related textual and/or visual domain-specific languages.
‒ Domain-specific languages facilitating to define the different aspects of mobile applications: user interface, data layer, static structure, dynamic behavior and communication protocols.
‒ A methodology that supports the common application of the different domain-specific languages.
Supporting the Development for Different Mobile Platforms and IoT Devices with Model Processing
‒ A model processing methodology that supports different mobile platforms at the same time.
‒ The correctness of the model processing algorithm has been proven. The produced artifacts are quality outputs and fulfill the requirements. Measurements verify the efficiency of the methodology.
Methods for Supporting Application in Environments with Limited Resources
‒ Investigation of those software and hardware factors that have effect on energy consumption in current mobile devices. The results have been analyzed.
‒ Based on the results of the analysis, development methodologies (best practices) and design patterns have been worked out. These results target the environments with limited resources, for example mobile devices.
Applying Cloud-based Technologies
‒ A method and related algorithms that applies cloud-based technologies for supporting domain-specific modeling.
‒ A method that applies cloud-based technologies for supporting model processing.
To summarize, the objectives have targeted one of the most pressing problems of mobile software development, which derives mainly from the diversity of mobile platforms. To solve this problem, the mobile platforms have been analyzed from different perspectives. A modeling language family has been designed for mobile applications and an optimal framework has been worked out for all platforms supporting the code generated from the models. The result is a complete methodology and a system that allows designing mobile
applications and with model processing effectively supports the application development for each major mobile platform.
This thesis shows that my research results, the continuous work in the field of information and communication systems, and the results of my group (Applied Informatics Group at the Department of Automation and Applied Informatics) from the last 15 years have a great impact in several fields and promoted several results: among others, the motivation of the PhD students, active research work of my colleagues, close relationships with several industrial partners, successful R&D projects.