Comparison results of different algorithm tests.

With the development of digital health, enhancing decision-making effectiveness has become a critical task. This study proposes an improved Artificial Bee Colony (ABC) algorithm aimed at optimizing decision-making models in the field of digital health. The algorithm draws inspiration from the dual-layer evolutionary space of cultural algorithms, combining normative knowledge from the credibility space to dynamically adjust the search range, thereby improving both convergence speed and exploration capabilities. Additionally, a population dispersion strategy is introduced to maintain diversity, effectively balancing global exploration with local exploitation. Experimental results show that the improved ABC algorithm exhibits a 96% convergence probability when approaching the global optimal solution, significantly enhancing the efficiency and accuracy of medical resource optimization, particularly in complex decision-making environments. Integrating this algorithm with the Chat Generative Pre-trained Transformer (ChatGPT) decision system can intelligently generate personalized decision recommendations and leverage natural language processing technologies to better understand and respond to user needs. This study provides an effective tool for scientific decision-making in digital healthcare and offers critical technical support for processing and analyzing large-scale medical data.