Scientific Workflow Clustering Based on Motif Discovery

Abstract

In this paper, clustering of scientific workflows is investigated. It proposes a work to encode workflows through workflow representations as sets of embedded workflows. Then, it embeds extracted workflow motifs in sets of workflows. By motifs, common patterns of workflow steps and relationships are replaced with indices. Motifs are defined as small functional units that occur much more frequently than expected. They can show hidden relationships, and they keep as much underlying information as possible. In order to have a good estimate on distances between observed workflows, this work proposes the scientific workflow clustering problem with exploiting set descriptors, instead of vector based descriptors. It uses k-means algorithm as a popular clustering algorithm for workflow clustering. However, one of the biggest limitations of the k-means algorithm is the requirement of the number of clusters, K, to be specified before the algorithm is applied. To address this problem it proposes a method based on the SFLA. The simulation results show that the proposed method is better than PSO and GA algorithms in the K selection.

Cross-domain graph based similarity measurement of workflows

“…A vast number of workflow representations have been designed for similarity measurement of workflows given about feature vectors [17]. However, workflow representation by graphs has several advantages over feature vectors.…”

“…Some workflow similarity measurements are text based [19], and some others consider different aspects of workflows (such as workflow motifs [17], or workflow structures [1,20]). For example a workflow similarity measurement method is the work at [19] that used feature selection techniques based on text.…”

“…We use workflow motifs [17], to enable relating workflows from one domain to other workflows by different domains. So far techniques were focused on discovering workflow motifs in one domain [17,[24][25][26]. The present paper addresses extraction of workflow motifs at intersection of different domains causing large amount of data.…”

“…Traditionally, tasks such as classification and clustering of workflow are solved by defining an applicable distance measure (such as Cosine, or Euclidean [17]) and then using an algorithm that is based on distances exclusively. Recently, a prominent alternative class of graph embedding methods based on spectral clustering has been used [12].…”

A Cloud Based Architecture for Working on Big Data with Workflow Management

“…Managers often reuse and refine workflows or current patterns of processes [1,2]. They can share their workflows or create new workflows from a public repository or a new project space (the extracted data, new components, sub-workflows, etc).…”