Design of improved JAYA algorithm for cigarette finished product logistics delivery

The accelerated proliferation of cigarette sales has precipitated an escalation in the industry's logistics distribution demand. This has given rise to salient issues, including elevated distribution costs and diminished delivery efficiency. These factors impede the enhancement of service quality and the fortification of market competitiveness. In response to these challenges, this study proposes an improved Jaya algorithm that integrates a reverse learning mechanism and a cosine similarity strategy to enhance optimization performance. The study also constructs a logistics delivery path planning model for finished cigarette products based on the enhanced algorithm. Experimental results demonstrated that the improved Jaya algorithm achieved approximately 40%-60% faster convergence compared to the traditional Jaya algorithm and the dynamic segmentation-based sine cosine algorithm across multiple benchmark test functions. In real-world cigarette logistics distribution scenarios, the improved algorithm reduced total distribution mileage by approximately 21%, lowered fuel consumption by about 29%, shortened delivery time by approximately 16%, and increased on-time delivery rates by 7%-12%. These results indicate that the proposed algorithm exhibits superior optimization capability, efficiency, and dynamic adaptability, providing strong support for enhancing the operational competitiveness of the cigarette logistics industry.