... and continuous reward generators – this method enables efficient learning in sparse reward scenarios. Moreover, a fuzzy reward function is designed based on Dempster–Shafer fusion theory, which reduces the complexity of designing traditional continuous reward functions and provides further efficiency support...