改进的分布估计鲸鱼算法求解区块链DAG任务调度问题

Improved distribution estimation whale algorithm forblock chain DAG task scheduling

为了克服区块链单链技术效率低的问题,一种新的范式有向无环图正在蓬勃发展。针对区块链有向无环图中考虑代价权重的非独立任务调度问题,构建了区块链DAG的任务调度数学模型,并为了求解该问题提出了一种基于改进分布估计鲸鱼的新任务调度算法。新算法在WOA中引入EDA的空间采样和统计学习来预测搜索的最佳区域,进而产生优秀的新个体,从而使得新算法具备更强的全局搜索能力和更快的收敛速度。最后通过程序仿真,对比了多种算法在收敛速度和全局寻优能力方面的性能。实验表明IEWOA比传统的WOA和EDA在以上参数性能有明显优势,相比于改进遗传算法FPGA,IEWOA同样在参数性能上有一定的优势。

In order to overcome the efficiency issues of single chain technology in blockchain,a new paradigm directed acyclic graphs is flourishing.This article focused on the non independent task scheduling problem considering cost weights in blockchain directed acyclic graphs,and constructed a mathematical model for task scheduling in blockchain DAG,and proposed a new task scheduling algorithm based on improved distribution estimation whale to solve this problem.The new algorithm introduced spatial sampling and statistical learning of EDA in the WOA algorithm to predict the optimal search area,thereby gene-rating excellent new individuals,making the new algorithm have stronger global search ability and faster convergence speed.Finally,through program simulation,it compared the performance of multiple algorithms in terms of convergence speed and glo-bal optimization ability.The experiments show that IEWOA has significant advantages over traditional WOA and EDA in the above parameter performance.In addition,comparing with the improved genetic algorithm FPGA,IEWOA also has certain advantages in parameter performance.