A Data-Driven Optimization Approach to Improve the Resilience of the Singapore Mass Rapid Transit Network

Improving the resilience of the Singapore Mass Rapid Transit network (SMRTN) requires both proactive robustification of the network and reactive emergency response planning in cases of disruptions. In light of the SMRTN expansion, the remodelled hub-and-spoke bus network, and new bridging services such as ridesharing services and automated driving vehicles, building an integrated resilience improvement plan for the SMRTN has become more complex and requires a systematic and tractable methodology. This work proposes to improve the resilience of the SMRTN through strategies obtained from a novel data-driven optimization approach that lays the foundations for cross-disciplinary work among three fully-fledged fields of research, namely, data analytics, optimization and simulation. It aims to use the available EZ-Link dataset at SMART to deliver an open-source software that is easily modifiable with new data, network topologies, and bridging services. The proposed optimization methodology is based on ambiguity-aversion, a method which uses signals from data to construct several legitimate interpretations of that data. Along the ambiguity-aversion line, the project proposes a collaboration of four modules: 1) a Data Analytics module, headed by SMART, to identify and quantify signals from available data, 2) an Uncertainty Quantification module, driven by BEARS, to construct uncertainty sets from signals 3) a Bridging Services module, headed by TUMCREATE, to study the flow of bridging services and to simulate and validate optimal solutions, and 4) an Optimization module, headed by SEC, to build an optimization model for integrated resilience improvement with data and analyses from other modules. This proposal intends to seed an Intra-CREATE Transportation Group for cutting-edge research on the resilience of the Singapore transportation system.