The project approaches important software engineering problems, its goal being the development of novel methods for automating different activities related to the maintenance and evolution of software using artificial intelligence techniques. The problems we have decided to approach are of major importance during software development and are complex, which is why computational intelligence algorithms were designed to find good solutions to them.

Objectives The major objective of this project is to contribute to improving the development process of software systems and their quality by obtaining innovative results in the search-based software engineering domain. An integrated software system will be also developed, with the purpose of assisting software developers in activities such as: defect prediction, package-level software refactoring and the identification of hidden dependencies in software systems. The present project is both applicative and interdisciplinary, having the following scientific objectives:

1. The development and experimental validation of innovative machine learning based methods for solving problems of major practical importance from the domain of software maintenance and evolution (the problems mentioned above).
2. The theoretical validation of the previously mentioned methods by elaborating new mathematical models for the problems considered.
3. The development and validation of the AMEL system - an integrated software system for the assistance of software developers in software maintenance and evolution activities (defect prediction, package-level software refactoring, identification of hidden dependencies).
4. Contribution to the knowledge development through the dissemination of the scientific theoretical and applicative results by publication in prestigious journals and participation to scientific conferences.

Original elements The project aims to bring original contributions in the directions of development of machine learning methods for important problems related to software maintenance and evolution. The original results obtained in the field of the project so far collectively by the members of the team are a confirmation of the fact that the research directions proposed in the present project are extremely promising, creating the premises for obtaining valuable scientific results. Therefore, we estimate that the original elements aimed by the present project will have a significant impact and importance in the development of the scientific knowledge in the domain of search-based software engineering. The estimated impact of the present project is presented as follows:

The scientific impact. The project targets, on the one hand, the theoretical fundamentals of the problems related to the adaptive optimization of the software systems using machine learning techniques. Considering that our purpose is the development of innovative computational models for defect prediction, package-level software refactoring and hidden dependencies identification, the research that we will carry out will have an important impact on the theoretical research of search-based software engineering problems.

Technological impact. The problems approached by this project derive from practical necessities of the software engineering field, therefore the approaches we propose in the present project might open the possibility of developing some tools/software technologies to be actually used by software companies. A software system such as AMEL might assist software developers in their daily maintenance and evolution tasks such as: software defect prediction, package-level refactoring and hidden dependencies identification. The solid mathematical foundations of the targeted approaches will enable the usage of the methods we wish to develop for software systems in critical domains (medicine, military, banking, etc.) as well.

Economical impact. The existence of a software system such as AMEL for actual software systems (developed and used in the industry) would be very useful since it would lead to a more efficient software evolution and maintenance process and to the reduction of maintenance costs.