DAG -extended deletion algorithm in Graphical Abstract Grid Workflow Model for remote sensing quantitative retrieval

Directed Acyclic Graph (DAG) has been widely used in Grid workflow modeling, since it deals with the mass data with a specific user-defined Grid Wordflow job scheduling. However, DAG has limit modeling power, for a DAG model cannot express process state information. DAG-extended model provides solutions to “glue” multiple algorithm attributes, scalable algorithm run-time environments, various matchmaking agents, parallel data features and operational attributes together. Currently, it is still a pending issue for the existing Grid workflow based on the DAG-extended scheduling to express complex relationships among various tight-coupling remote sensing algorithms with a series of definition toolkit, such as deletion algorithm, for it can only describe the logical feature of remote sensing processing algorithms but can not show the parallel feature of algorithms. Aiming at this problem, this paper proposes a DAG -extended deletion algorithm in Graphical Abstract Grid Workflow Model for remote sensing quantitative retrieval application. In this paper, we mainly: (1) discuss the limits of current Grid workflow models applied in the remote sensing field, and (2) modify the traditional DAG Grid Workflow model to the DAG-extended Grid Workflow model, especially describing the logical feature of remote sensing algorithms. Besides, (3) based on the new defined tool - DAG-Extended deletion algorithm in graphical Grid workflow model, we give the concrete implementation example to present the dynamic modification achievement of this composition tool and illustrate its benifits over algorithms based on the traditional DAG Graphic-oriented Abstract Grid Workflow model.