System architecture for information visibility in humanitarian logistics: innovative lateral information sharing in a low trust, high risk, volatile environment for coordination
Cameron Taylor · ResearchSpace (University of Auckland) · 2019
Interagency coordination is one of the most enduring and intractable problems affecting humanitarian logistics performance in disaster relief. This thesis seeks to understand the reasons and solutions to poor interagency performance. A literature review surveys seminal and recent research for supply chain coordination issues in disaster relief. A primary driver of supply chain coordination is information sharing, which requires information visibility. Many supply chain strategies and coordination mechanisms in the traditional literature are inappropriate for disaster relief. By integrating supply chain theory into design science, the combined framework outlines the research plan to prototype a system for information sharing. System dynamics simulations with volatile demand are used to test the effects of different information sharing scenarios across multiple supply chains on backlogs. Results show horizontal information sharing between competing retailers is as effective as other scenarios after an initial 18 day delay, as orders filter upstream to suppliers. Select supply chain theories are turned into design principles and a conceptual framework, which is then instantiated as a generic system architecture and practical artefact. The practical artefact functions by using autonomous decentralised mesh networks to facilitate an automated data collection and anonymisation application, which forwards securely anonymised and semi-aggregated data through an information sharing process called "lateral information sharing". This allows supply chain competitors to share information without needing to trust each other, but still gaining the benefits of information sharing as if they were in a strategic partnership. This has limited generalisability beyond disaster relief due to competition laws, while security issues around implementation together with trust issues around adoption will determine whether this approach is feasible beyond a laboratory scenario.