Network Reliability

Network reliability is the most popular index for evaluating system performance. In general, network reliability is defined as the probability of success for all commodities in sending a given amount of data/flow/information from source nodes to sink nodes under special conditions. In this project, we also aim to exploit a variety of new algorithms for system modeling and reliability index calculation in traditional networks and MMNs, universal generating function methods (UGFM), and Soft Computing.

Networks have broadly modeled real-world systems such as cloud computing, supply chain systems, electrical power or transportation networks, etc. Hence, networks plays an important role in modern society and has consequently attracted much study attention. Reliability is one of the most popular network performance indexes.

The escalating proliferation of data during the current Big Data Era has resulted in data that is high-volume, high-velocity, and/or high-variety. As a result, current systems such as supply chain management tend to be too complex to model and evaluate using only traditional network methodologies. In this project, we aim to develop a novel network called the multi-configuration multi-state network (MMN) to model, analyze, and design complex systems. The MMN is competent at integrating multi-state, multi-attribute components (e.g., cost, bandwidth, capacity) and multi-characteristic flow structures to allow commodities such as tasks, files and video streams to be fully functional within their own configurations, thereby surpassing the capabilities of traditional methodologies.

Our research is mainly based on the areas of network reliability and focuses on bridging the gap between cutting edge theoretical research and challenging real-life problem-solving. Our research have motivated and inspired researchers to work on reliability areas, including complex network analysis and Soft Computing based reliability evaluation methodologies. The algorithms (MP/MC/d-MP/d-MC) that we proposed for flow network reliability, whether binary-state or multi-state, single commodity or multiple commodities, are presently the top in the field from published literature records. Further research is being undertaken to expand the scope of accessibility into new fields, such as logistics management and multimedia.