Journal article
Achieving High Robustness in Supply Distribution Networks by Rewiring
IEEE transactions on engineering management, Vol.58(2), pp.347-362
05/01/2011
DOI: 10.1109/TEM.2010.2095503
Abstract
In this paper, we propose a new rewiring approach for distribution networks called randomized local rewiring (RLR). We evaluate the robustness of original and rewired distribution networks using new metrics and show that the choice of a network topology can affect its robustness considerably. Some supply and distribution networks exhibit characteristics similar to those of scale-free networks. Simulation results show that applying RLR to such distribution networks can improve the network robustness on the supply availability and network connectivity metrics when both random and targeted disruptions are likely to occur. A unique feature of our model is a tunable rewiring parameter, which makes it possible to design networks with the same performance on the supply availability, network connectivity, and average delivery efficiency metrics in the presence of both types of disruptions. This paper will describe the robustness metrics and the new approach, illustrate the experimental results in the context of a military logistic and a retailer's distribution network, and discuss the insights gained about choosing the right topology for achieving higher robustness.
Details
- Title: Subtitle
- Achieving High Robustness in Supply Distribution Networks by Rewiring
- Creators
- Kang Zhao - Pennsylvania State UniversityAkhil Kumar - Pennsylvania State UniversityJohn Yen - Pennsylvania State University
- Resource Type
- Journal article
- Publication Details
- IEEE transactions on engineering management, Vol.58(2), pp.347-362
- Publisher
- IEEE
- DOI
- 10.1109/TEM.2010.2095503
- ISSN
- 0018-9391
- eISSN
- 1558-0040
- Number of pages
- 16
- Grant note
- Smeal College of Business, Pennsylvania State University Pennsylvania State University
- Language
- English
- Date published
- 05/01/2011
- Academic Unit
- Business Analytics
- Record Identifier
- 9984380477302771
Metrics
2 Record Views