当前位置: X-MOL 学术Genet. Program. Evolvable Mach. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
SonOpt: understanding the behaviour of bi-objective population-based optimisation algorithms through sound
Genetic Programming and Evolvable Machines ( IF 2.6 ) Pub Date : 2023-03-13 , DOI: 10.1007/s10710-023-09451-5
Tasos Asonitis , Richard Allmendinger , Matt Benatan , Ricardo Climent

We present an extension of SonOpt, the first ever openly available tool for the sonification of bi-objective population-based optimisation algorithms. SonOpt has already introduced benefits on the understanding of algorithmic behaviour by proposing the use of sound as a medium for the process monitoring of bi-objective optimisation algorithms. The first edition of SonOpt utilised two different sonification paths to provide information on convergence, population diversity, recurrence of objective values across consecutive generations and the shape of the approximation set. The present extension provides further insight through the introduction of a third sonification path, which involves hypervolume contributions to facilitate the understanding of the relative importance of non-dominated solutions. Using a different sound generation approach than the existing ones, this newly proposed sonification path utilizes pitch deviations to highlight the distribution of hypervolume contributions across the approximation set. To demonstrate the benefits of SonOpt we compare the sonic results obtained from two popular population-based multi-objective optimisation algorithms, Non-Dominated Sorting Genetic Algorithm (NSGA-II) and Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D), and use a Multi-objective Random Search (MRS) approach as a baseline. The three algorithms are applied to numerous test problems and showcase how sonification can reveal various aspects of the optimisation process that may not be obvious from visualisation alone. SonOpt is available for download at https://github.com/tasos-a/SonOpt-2.0.



中文翻译:

SonOpt:通过声音理解基于种群的双目标优化算法的行为

我们介绍了 SonOpt 的扩展,这是第一个公开可用的工具,用于双目标基于种群的优化算法的超声化。SonOpt 已经通过提议使用声音作为双目标优化算法过程监控的媒介,引入了对算法行为理解的好处。第一版 SonOpt 使用两种不同的声化路径来提供有关收敛、种群多样性、连续几代目标值的重复以及近似集形状的信息。本扩展通过引入第三条超声化路径提供了进一步的见解,该路径涉及超体积贡献以促进对非支配解决方案的相对重要性的理解。使用与现有方法不同的声音生成方法,这种新提出的声化路径利用音高偏差来突出近似集上超体积贡献的分布。为了证明 SonOpt 的优势,我们比较了从两种流行的基于种群的多目标优化算法、非支配排序遗传算法 (NSGA-II) 和基于分解的多目标进化算法 (MOEA/D) 获得的声波结果,并使用多目标随机搜索 (MRS) 方法作为基线。这三种算法应用于众多测试问题,并展示了可听化如何揭示优化过程的各个方面,这些方面仅通过可视化可能并不明显。SonOpt 可在 https://github.com/tasos-a/SonOpt-2.0 下载。这个新提出的超声化路径利用音高偏差来突出超体积贡献在近似集上的分布。为了证明 SonOpt 的优势,我们比较了从两种流行的基于种群的多目标优化算法、非支配排序遗传算法 (NSGA-II) 和基于分解的多目标进化算法 (MOEA/D) 获得的声波结果,并使用多目标随机搜索 (MRS) 方法作为基线。这三种算法应用于众多测试问题,并展示了可听化如何揭示优化过程的各个方面,这些方面仅通过可视化可能并不明显。SonOpt 可在 https://github.com/tasos-a/SonOpt-2.0 下载。这个新提出的超声化路径利用音高偏差来突出超体积贡献在近似集上的分布。为了展示 SonOpt 的优势,我们比较了从两种流行的基于种群的多目标优化算法、非支配排序遗传算法 (NSGA-II) 和基于分解的多目标进化算法 (MOEA/D) 获得的声波结果,并使用多目标随机搜索 (MRS) 方法作为基线。这三种算法应用于大量测试问题,并展示了可听化如何揭示优化过程的各个方面,这些方面仅通过可视化可能并不明显。SonOpt 可在 https://github.com/tasos-a/SonOpt-2.0 下载。

更新日期:2023-03-13
down
wechat
bug