In an effort to improve performance of congested gateways, a new Active Queue Management (AQM) algorithm, Adaptive Virtual Queue Random Early Detection (AVQRED), was developed by feeding virtual queue size to the RED algorithm. The objective of the new algorithm is to improve overall performance by keeping link utilization high, link utilization stable, queueing delay low and consecutive packet drop rate low. This paper shows the objective is met by comparing AVQRED with six other well known AQM methods in a realistic emulation environment. To provide fair comparisons, the AQM parameters are fine-tuned by exploring many different parameter settings via emulations. The emulation results conclude that AVQRED improves overall performance by 8 to 25%. To provide intuitive properties and validate the emulation results, a mathematical model is proposed and fed to MATLAB, and the results from MATLAB and emulations are compared.