This short review is based on the article Caputo et al. (Inverse Problems 35:055006, 2019). We introduce a numerical algorithm to detect a source in a network carrying a miscible flow and apply it to a fourteen vertex graph to understand what are the numerical issues involved. We found that the least square estimation should be done using at least two vertices and that localized eigenvectors of the Laplacian can cause difficulties. Assuming this estimation is done, the source detection algorithm works well except for some singularities coming from the structure of the graph.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Source Detection on A network: Numerical Considerations

  • Jean-Guy Caputo

摘要

This short review is based on the article Caputo et al. (Inverse Problems 35:055006, 2019). We introduce a numerical algorithm to detect a source in a network carrying a miscible flow and apply it to a fourteen vertex graph to understand what are the numerical issues involved. We found that the least square estimation should be done using at least two vertices and that localized eigenvectors of the Laplacian can cause difficulties. Assuming this estimation is done, the source detection algorithm works well except for some singularities coming from the structure of the graph.