Tracking Dynamic Objects using Opposition-Based Dierential Evolution
Fadi Salama · 2007
Recently, many new approaches have been developed and applied to the problem of tracking dynamic objects. Some of these approaches include the use of dynamic optimization techniques. This project focuses on applying a dynamic optimization technique to the problem of tracking dynamic objects. To do this, different optimization algorithms were compared and the best approach was selected and implemented. The pool of selection included: evolutionary algorithms, differential evolution and particle swarm optimization techniques, with the successful candidate being differential evolution. A particular strand of differential evolution, known as Opposition-based Differential Evolution, along with three different strategies, namely ‘Dispersing Converged Population Members (DCPM)’, ‘Improved Comma Strategy’, and ‘BestWorst’ were implemented. These were tested on a number of different experimental setups and their performance were recorded and analyzed. The new strategies developed were found to perform better than ODE in tracking dynamic objects.