A Context-Aware Approach Framework to Algorithm Selection for Search Based Procedural Content Generation

This study challenged the idea that a single optimal algorithm exists for Procedural Content Gener-ation. This emphasises the crucial role of task-specific considerations in determining the most appropriate al-gorithm for a given task. Through a novel comparative analysis and the results of our previous studies, we demonstrate the importance of contextual superiority. PSO has emerged as highly effective in generating race tracks, whereas GA has demonstrated superior performance in map generation. This study extends the traditional focus on GAs by highlighting the unique strengths of PSO and ABC. This advocates broader exploration beyond the confines of GA-centric approaches to PCG. We propose an approach termed “Context-Aware Ap-proach” for algorithm selection, which underscores the importance of analyzing the complexity of the task, qual-ity requirements, and desired solution speed. By adopting this context-aware approach and comprehending the interactions between task characteristics and algorithm capabilities, researchers and developers can make well-informed decisions, ultimately leading to the creation of engaging and personalized PCG experiences.