Automatizace tvorby programů v prostředí LabVIEW pomocí genetických algoritmů s využitím prvků umělé inteligence

Abstract

This dissertation deals with the issue of automatic code generation, based on user-defined requirements. The aim was to use artificial intelligence algorithms, specifically evolutionary algorithms, to automatically create parts of programs so that the outputs of the created algorithm fully correspond to the requirements of the user, who does not need to have any basic programming knowledge in this case.

In the theoretical part, the basic assumptions and principles of finding optimal solutions using genetics, genetic programming and evolutionary algorithms are summarized. The principles of operation in the real world and their subsequent transfer to mathematical models and the world of information technology are contained. The work also provides a mathematical proof of the convergence of partial solutions of problems using genetic programming.

The actual automatic code generation was performed in the LabVIEW programming environment. The work therefore further deals with the main idea of the operation of this programming environment and describes why LabVIEW is a suitable tool for solving this issue. The practical part of the work then describes the proposed concept suitable for achieving the desired results and applies the theoretical procedures mentioned.

The final part of the work is devoted to the validation of the acquired results, specifically for automatic code generation based on user-defined inputs and their desired outputs for Boolean, Numeric and String system variables. At the same time, this issue is extended to their mutual combinations and cycles. The resulting part of the work is devoted to the results when generating sequences.