Cognitive computing has been commonly used to address different forms of optimization issues. Swarm intelligence (SI) and evolutionary computing (EC) are population-based intelligent stochastic search techniques promoted to search for their food from the intrinsic way of bee swarming and human evolution. Initialization of populations is a critical factor in the Particle swarm optimization (PSO) algorithm that significantly affects diversity and convergence. Quasi-random sequences based on cognitive computing are more helpful in initializing the population than applying the random distribution for initialization to maximize diversity and convergence. The capacity of PSO is expanded to make it suitable for the optimization problem by adding new initialization techniques based on cognitive computing using the sequence of low discrepancies. The employed low discrepancies sequences included WELL named WE-PSO to solve the optimization problems in large-scale search spaces. The proposed approach has been tested on fifteen well-known uni-modal and multi-modal benchmark test problems extensively used in the literature. Also, WE-PSO efficiency has been compared to standard PSO, and two other Sobol-based PSO (SOB-PSO) and Halton-based PSO (HAL-PSO) initialization approach. The results were obtained to validate the efficiency and effectiveness of the proposed approach. Mean fitness values obtained using WE-PSO designate that WE-PSO is better than standard techniques in multi-modal problems. The computational results also show that the proposed technique outperformed and has a higher accuracy rate than the classical approaches. Besides, the proposed work’s result offers a foresight of how the proposed initialization approach has a significant effect on the importance of cost function, convergence, and diversity.

认知计算已普遍用于解决不同形式的优化问题。群智能（SI）和进化计算（EC）是基于人口的智能随机搜索技术，被推广为从蜂群和人类进化的内在方式中搜索食物。种群初始化是粒子群优化（PSO）算法中的一个关键因素，它会极大地影响多样性和收敛性。与应用随机分布进行初始化以最大程度地提高多样性和收敛性相比，基于认知计算的准随机序列对初始化总体更为有用。通过添加基于低差异序列的基于认知计算的新初始化技术，PSO的容量得以扩展，使其适用于优化问题。为了解决大规模搜索空间中的优化问题，所采用的低差异序列包括名为WELL的WE-PSO。所提出的方法已经在文献中广泛使用的十五个众所周知的单峰和多峰基准测试问题上进行了测试。此外，已将WE-PSO效率与标准PSO以及其他两种基于Sobol的PSO（SOB-PSO）和基于Halton的PSO（HAL-PSO）初始化方法进行了比较。获得的结果证实了所提出方法的效率和有效性。使用WE-PSO获得的平均适应度值表明，在多模式问题中，WE-PSO优于标准技术。计算结果还表明，所提出的技术优于经典方法，并且具有更高的准确率。除了，