This paper proposes a dynamic ring architecture of interaction among members of population in a Quantum Evolutionary Algorithms (QEA). The ring is allowed to expand/collapse based on a sinusoidal population size and partial re-initialization of new members in the population. The study shows that the ring structure can be an efficient architecture for an effective Exploration/Exploitation tradeoff, and the partial reinitialization of the proposed algorithm can improve the diversity of the algorithm. The proposed approach is tested on Knapsack Problem, Trap Problem as well as 14 numerical optimization functions. Experimental results show that the proposed Structure consistently improves the performance of QEA.

Keywords