In the era of advanced nanotechnology where billions of transistors are fabricated in a single chip, high-speed operations are challenging due to packaging related issues. In High-Speed Very Large Scale Integration (VLSI) systems, decoupling capacitors are essentially used in power delivery networks to reduce power supply noise and to maintain a low impedance of the power delivery networks. In this paper, the cumulative impedance of a power delivery network is reduced below the target impedance by using state-of-the-art metaheuristic algorithms to choose and place decoupling capacitors optimally. A Matrix-based Evolutionary Computing (MEC) approach is used for efficient usage of metaheuristic algorithms. Two case studies are presented on a practical system to demonstrate the proposed approach. A comparative analysis of the performance of state-of-the-art metaheuristics is presented with the insights of practical implementation. The consistency of results in both the case studies confirms the validity of the proposed appraoch.

中文翻译：

---

使用基于矩阵的进化计算有效选择和放置封装内去耦电容器

在单个芯片中制造数十亿个晶体管的先进纳米技术时代，由于封装相关问题，高速操作具有挑战性。在高速超大规模集成 (VLSI) 系统中，去耦电容器主要用于供电网络，以降低电源噪声并保持供电网络的低阻抗。在本文中，通过使用最先进的元启发式算法来最佳地选择和放置去耦电容器，将供电网络的累积阻抗降低到目标阻抗以下。基于矩阵的进化计算 (MEC) 方法用于有效使用元启发式算法。在一个实际系统上介绍了两个案例研究，以证明所提出的方法。对最先进元启发式算法的性能进行了比较分析，并结合了实际实施的见解。两个案例研究结果的一致性证实了所提议方法的有效性。