Marc Wouters

2024

The automotive industry is undergoing a fundamental transformation. In the future, cars will be defined more by their software than by their hardware. With the increasing amount of software in vehicles, and especially the rising share of software costs, software cost estimation is becoming a critical component of planning for automotive manufacturers. This highlights the growing importance of reliable and accurate software cost estimation methods for the automotive industry. When examining the literature, there is surprisingly little research focused on software cost estimation practices in the automotive industry, despite its unique development environment and the enormous complexity of automotive software. This is surprising, given that the automotive industry presents a unique environment for software development, and the complexity of automotive software is incredibly high. In this systematic mapping study on software cost estimation practices in general, a large portion of the literature is analyzed and classified. I analyze the existing literature regarding its research intensity on the topic of software cost estimation practices. Furthermore, I examine the methodological approach in terms of research type, research method, and contribution type. In the next step, I investigate the published model proposals in more detail, focusing on their diversity, different categories, and the datasets used for testing. Finally, I compile these results into an appropriate map to identify future research gaps and directions. My findings suggest a very one-sided and flawed approach in the literature, which, in the long term, will not be effective in identifying new models for practical application, particularly in the automotive industry.