SIMTech-NTU Joint Lab on Complex Systems
Block N4, B1a-01 | School of Computer Engineering, Nanyang Technological University | Singapore 639798

News


Agency for Science, Technology and Research (A*STAR) funded TSRP on Logistics & Supply Chain Management Programme: Multi-Objective Vehicle Routing for Last Mile Logistics
28 January 2015 - 1:33pm

In this project, Associate Professor One Yew Soon, Director of the Centre for Computational Intelligence at the School of Computer Engineering at Nanyang Technological University, and his team collaborate with Dr. Tan Puay Siew, Manager of Production & Operations Management group at Singapore Institute of Manufacturing Technology (SIMTECH), and her team to develop multi-objective and dynamic vehicle routing for last mile logistics in city area. This research is also in the interdisciplinary collaboration with Associate Professor Mark Goh, Assistant Professor Cheong Taesu, Dr. Robert de Souza at The Logistics Institute Asia Pacific (TLIAP), National University of Singapore and Associate Professor Lau Hoong Chuin at Singapore Management University. The last mile logistics is currently regarded as one of more expensive, least efficient and most polluting sections of the entire logistics chain in an urban environment. This project tackles the challenges by developing innovative technologies for freight planning and scheduling for the urban last mile logistics. This project aims to answer the following research challenges:

  • How to maximise the utilisation of commercial traffic servicing retailers and other business users at downtown to increase quality of city life?
  • How to reduce the amount of heavy goods commercial delivery vehicles to reduce congestion and improve air quality?
  • How to relieve pressures on parking and loading/unloading areas in city centres or malls?

In this project, we seek to improve the efficiency and reduce environmental impact for the urban last mile logistics that can be modelled as a complex logistics system consisting of clusters of customers, suppliers, and logistics service providers (LSPs) interacting through a marketplace (i.e. E-Market) as shown in Figure 1. To do so, we propose to develop Multi-objective Vehicle Route Planning (MoVRP) algorithms to help LSPs plan and optimise their logistics services for common routes with multiple pickup and delivery points according to the schedules from the E-Market, with environmental impact considerations in the complex urban last mile logistics system. In addition, Dynamic Routing and Scheduling algorithms that enable LSPs to perform on-the-fly vehicle routing and scheduling of vehicles based on real-time traffic conditions for common route with multiple pickup and delivery points and repair a route to any changes in environmental/traffic conditions. Above the standard objectives of improving logistics services, eco-indicators are taken into consideration as additional objectives or constrained variables when helping LSPs plan and optimise their logistics services. Particularly, the joint lab will embarked on the identifications of eco-indicators and the development of model cum algorithms for the quantification of these eco-indicators (QEI) based on multi-factor such as relationships between vehicle types, tonnage/loading, road conditions, traffic conditions, fuel consumptions and emissions, etc. Case studies will be conducted and verified with selected relevant industry collaborator(s).


Figure 1

First, in response to the strong demand and environmental challenges for last mile logistics, we propose to build on our technological innovations in planning and scheduling, dynamic vehicle routing and estimation of the carbon footprint to create a novel technology for eco-friendly vehicle planning and dynamic routing for last mile logistics in an urban environment. Our proposed work to develop novel algorithms for vehicle route planning and dynamic vehicle routing and scheduling in consideration of environmental impact and collaboration through E-Market will extend conventional VRP research to a new level of complexity as new dimensions are added with the introductions of eco-indicators. Second, a multi-objective VRP formulation with eco-indicators from a complex science perspective will provide a new insights and solutions that are more tractable and hence better equipped to solve multi-faceted challenges in last mile logistics, with specific considerations of urban constraints. Lastly, while there is a range of separate works available in the individual areas of fuel consumption estimation modelling, emissions modelling and impact assessment modelling, the suitability of their use in last mile logistics is indeterminate as there is no specific method to integrate and synergise the three components. With the proposed integrated approach, a seamless unified method can be potentially established in the evaluation of eco-indicators in support of the last mile logistics. This research targets to achieve the following milestones:

  • Costs and Time Savings: Logistics Cost: 10-30%, Planning Time: 40-50%
  • Quick Response & Adaptability: Collaboration allows easy outsourcing, Optimisation algorithms allows for efficient route, and planning, scheduling and repairs
  • Corporate Social Responsibility: Greener operations, with in-built cost-efficiency

Agency for Science, Technology and Research (A*STAR) funded TSRP on Master Facilitative Control Tower For Risk Management of Complex Supply Chains Logistics & Supply Chain Management.

On this premise, Associate Professor One Yew Soon, Director of the Centre for Computational Intelligence in the School of Computer Engineering at NTU, and his team collaborate with Dr. Tan Puay Siew, Manager of Production & Operations Management group at SIMTECH, and her team to analyse complex systems and develop risk management model in supply chain network, see Figure 2.


Figure 2

The increases in scale, connectivity and vulnerability make managing the complex dynamics in supply chains and manufacturing systems, such as bullwhip effect and disruptions, very difficult. Analytical based approaches are difficult to understand how a complex supply chain works as a whole and the interplay of various factors and components. The traditional models also assume the structure and components of supply chains are fixed and actually this is not true and supply chains are dynamic all the time. To breaks the limitations of traditional approach, the complex systems programme intends to provide a better choice on investigating the behaviors of supply chains and enables the investigation of emergent behaviors and issues in complex supply chain networks. A local control tower oversees the risk issues in the supply chain of a focal company. However, one control tower only covers the focal supply chain. In case of disruptions, it needs to collaborate with other control towers to reduce risks or enable mitigations through information or resource sharing. A master control tower will facilitate this collaboration and provide integrated and quick fix solution for managing the disruption. This research serves as a platform for encouraging active interdisciplinary collaboration between NTU and SIMTECH in the frontier research of Complex Systems in manufacturing operations and supply chain. We nurture Science & Engineering talents through manpower training for PhD students in the area of Complex System with the goal of educating PhD holders capable of independent work in a multidisciplinary environment. The project also proposes new concepts that may lead to new and more efficient ways of supporting industry needs in complex manufacturing and smart production operations and supply chain systems. A bottom up analysis approach of complex supply chain networks by complex systems approaches (agent-based modeling and evolutionary computation). The new approach overcomes limitation of traditional approaches on understanding how a complex supply chain works as a whole and what are the interplays of factors and components in a risk scenario. 'What-if' analysis of Disruptions and Mitigations This research proposes a two layer approach. In the Upper Layer, we produce the reference value of inventory position for a period time in future based on optimization of supply chain level performance. At the Low Layer, inventory position profile from the upper layers is used by the low layer control module (MPC controller) as reference for maintaining inventory position with considering actual demand. Two Layer approach for Mitigation with Evolutionary and Control Algorithms The technologies and models developed in the project intend to be beneficial to industries in meeting the following challenges:

  • growing supply chain complexities due to economic globalization
  • increasing supply chain risk and vulnerability
  • requirements on supply chain flexibility and responsiveness
  • needs for modeling supply chains as adaptive systems especially under the scenario of supply chain risk management

The technologies developed will be mainly applied to high-tech, aerospace, maritime and chemical industries, to help companies on solving complex supply chain problems. We expect that the technology and prototype have a high potential for industry scenarios where complex supply chains evolve over time, especially where complex scenarios, uncertainties and variations are highly expected. Potential companies: Big M&C companies and SMEs. E.g., IBM, Infineon, Maha, IDSC.