Robust optimization of chemical process networks based on Louvain-KBICD community division rewiring algorithm

Abstract

Previous work using the rewiring algorithm for robust optimization of chemical process networks did not take into account the existence of community structures between networks, thereby reducing the extent of robust optimization. Therefore, this paper proposes a robust optimization of chemical process networks based on the Louvain-KBICD community division rewiring algorithm. This algorithm firstly employs the K-shell-based algorithm with improved comprehensive degree (KBICD) to identify the key nodes of the network; it then proposes the community division of the network based on the Louvain-KBICD algorithm; and finally, it performs a robust optimization, respectively, by using the rewiring algorithm that reserves the node degree within the communities and the intelligent rewiring algorithm based on the average degree improvement between the communities. The case study proves that the key nodes identification algorithm proposed in this paper solves the problems of low resolution and insufficient identification accuracy of the previous algorithms, and the resolution is improved by 0.6607 and 0.8139 compared with the benchmark algorithm, respectively; the community division algorithm improves the quality of the network community division, and reduces the complexity of the community division, improving the quality of the community division by 11.20% and 14.58%, respectively; and the robust optimization algorithm effectively improves the extent of robust optimization of chemical process networks and preserves the initial community structure of the network while optimizing, meaning the robust optimization extent can reach 62.19%, 80.97% and 64.94%, 76.39% under two attacks, respectively.