Novel Blind Load Balancing Scheduling Algorithms for Distributed Tracking Networks
Thomas L. Frey, David K. Faulk · 2016
A novel algorithm is presented for quickly balancing responsibilities across a distributed tracking network that requires no explicit coordination between network members. Previous load balancing methods reliably achieved an equilibrium state; however, in times of large disparities (initialization) or discontinuities (loss of a participant), previous methods may require a significant time to reach the balanced state, perhaps several minutes, and experience significant overshoot before achieving a stable and balanced configuration. The improved algorithm offers substantial improvement in convergence time to equilibrium and eliminates the interim overshoot behavior of some previous algorithms.