Cloud-Based Low-Code Model Transformations Composition and Execution

Low-code development platforms offer a streamlined approach to software development, utilizing visual interfaces and drag-and-drop utilities instead of traditional programming languages. This allows for faster application development and deployment, granting non-technical users access to tailored software solutions based on their unique needs. One key aspect of low-code development is the composition of model transformations, which enables developers to combine and reuse pre-existing models and components to create new applications. Model transformation composition refers to the process of combining multiple, simpler transformations to achieve a desired outcome. These compositions can range from converting one model format to another, extracting information, or manipulating data within a model. In various industries, such as software engineering, business process management, and product design, the use of model transformation composition can automate the creation of new software systems, improve business processes, and aid in product design. The process of composing transformations presents a significant challenge to developers. Typically, smaller transformations are sourced from different and diverse sources and then manually combined, leading to a time-consuming and error-prone composition process. The application in low-code development platforms and the use of model transformation composition enable organizations to quickly and efficiently create and deploy business-critical applications based on the concept of model transformation and its composition. This thesis aims to streamline the process by solving the issue of chaining various model transformations to create complex models. The solution involves externally combining simpler transformation steps to achieve this goal. The first step is identifying the different transformations that need to be chained. After identifying these transformations, selecting only those specific to the user is crucial, considering quality criteria such as the metamodels and transformations used. A search-based optimization approach utilizing model-driven techniques has been employed to tackle this selection problem. This process leads to optimizing the execution of the selected transformation chains, reducing the number of generated target elements and improving the overall execution time. Thus, this thesis is focused on finding a solution to the complex problem of composing transformations by utilizing search-based optimization and model-driven techniques to identify, select, and optimize the execution of the most efficient transformation chains according to user requirements.