Fully Automated Game Testing via Neuroevolution

The video gaming industry thrives with an expected record revenue of $365.60 billion in 2023 and an annual growth rate of 6.52% from 2023 to 2027 [1] . To gain a foothold in this emerging market, developers have to ensure the best gaming experience possible, which can only be achieved via extensive testing procedures. However, most video games are created incrementally; some are even developed indefinitely, resulting in many program increments that have to be tested over and over again. Even though fully automated testing of games could relieve developers from this tedious task, a look at current industry practices reveals a dire need for more research, as most companies still test manually [2] , [3] . Besides entertaining the player, video games are also increasingly used for programming education because games keep the students motivated while demonstrating crucial programming concepts [4] – [6] . Although even experienced programmers rely on integrated development tools, students are left off with a lack of tools to assist them during their learning journey. Thus, more work on automated game testing is required such that students and practitioners cannot only validate the correctness of programs but also use generated inputs for dynamic program analysis.