Dynamic Load Distribution of Shortest-Path Finding in Client-Server Public Transit System
Yosua Raka Justico, Muhammad Yasir Anshari Haq, Aryo Pinandito · 2023
Mobile devices are very popular in the modern digital age and have many uses, one of which is that they play a crucial part in supporting navigation with specific navigation applications. This study addresses the challenges posed by varying smartphone device quality and the performance of the Dijkstra algorithm in navigation systems. It highlights the importance of efficient navigation applications, e.g., Waze and Google Maps, that employ algorithms such as Dijkstra to find the shortest routes. However, the performance of Dijkstra can vary depending on the smartphone's quality and system architecture, particularly the amount of data processed, impacting user satisfaction. To tackle these issues, dynamic load distribution mechanisms are proposed. These systems aim to optimize resource utilization by dynamically reallocating resources based on the complexity of calculations and device capabilities, offering a consistent and efficient navigation experience. The study's primary objective is to enhance Dijkstra's workload management, particularly in response to fluctuations in client-server loads. This approach ensures real-time adjustments to Dijkstra calculations, preventing server overloads, and maintaining a seamless navigation experience. The study's findings suggest that the hardware's memory capacity significantly influences Dijkstra's calculation time. Larger memory capacity results in quicker Dijkstra calculations, emphasizing the importance of high-memory devices for optimal performance. However, load balancing and task transfer to servers can effectively mitigate performance degradation when Dijkstra runs on lower-quality devices. Ultimately, finding a smooth navigation experience with a waiting time not exceeding one second is the goal of the study.