<p>In many practical applications of multi-component systems, the failure of one or more components redistributes the workload among the remaining units, leading to higher failure rates and accelerated degradation. This redistribution directly affects the overall reliability of the <i>k-out-of-n:G</i> system, potentially altering the number of components needed to successfully complete its mission. This paper examines the optimal redundancy allocation problem in a <i>k-out-of-n:G</i> system, considering both the system’s mission duration and the load distribution among its components. The system operates under two preventive replacement strategies, referred to as <i>First</i> and <i>Last</i> replacement. Under these strategies, the system is preemptively replaced either at a predetermined age <i>T</i> or upon completing a random mission duration <i>Y</i>, depending on whether the replacement occurs at the first or last of these events. To describe the effect of varying stress levels on the failure time of the system, we adopt the Cumulative Exposure (CE) model. We develop cost-effective design strategies for <i>k-out-of-n:G</i> systems, demonstrating that the optimal configurations computed using a proposed algorithm yield performance that is lower than or equal to that obtained under the assumption that neglects the load-sharing effect. Moreover, as the stress acceleration factor on the surviving components increases, the optimal design under the first strategy converges to a parallel system more rapidly than under the last strategy.</p>