<![CDATA[This study designs and implements a web-based sales information system for PT Putra Indo Mandiri Sejahtera by applying the Multilevel Feedback Queue (MLFQ) algorithm with dynamic time quantum to optimize transaction scheduling based on product urgency levels. The problem addressed is the absence of a transaction processing mechanism that considers product characteristics, causing fresh milk transactions with short shelf lives to be treated equally with processed products such as cheese. A quantitative method was employed with a population of 1,200 transactions and a sample of 300 transactions selected through stratified random sampling. The MLFQ system uses three queue levels mapped according to product shelf life: Q1 for fresh milk and yoghurt, Q2 for gelato, and Q3 for cheese. The dynamic time quantum was calculated from the absolute difference between the mean and median of transaction burst times, yielding tq1 = 141 seconds, tq2 = 282 seconds, and tq3 = 564 seconds. Simulation results show that high-urgency product transactions were processed earlier, while transactions with larger burst times were automatically demoted to the appropriate queue level. The implementation of MLFQ demonstrates improved transaction scheduling efficiency and proportional service quality.]]>
Copyrights © 2026
