Analysis and optimization of a logistics system to improve performance

• Main Author: Lim, Theresa Hui Si
• Other Authors: Appa Iyer Sivakumar
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/72072
• Institution: Nanyang Technological University

This project presents the study of the transportation operations of a third-party logistics (Singapore). Through performance analysis and modelling, the author seeks to optimize the current road transportation network through a mathematical model. The mathematical model developed aims to show improvement in performance for reducing logistics transportation costs. The literature review is conducted to gain an understanding of the relevant vehicle routing and scheduling problem (VRP) pertaining to allocation of transportation orders to a fleet of vehicles, arrangement of visiting cities / nodes for each vehicles, and in many alternative forms depending on the natural of operations, time window for decision making, and types of constraints that must be adhered to. In the present study, the problem involved 20 cities / nodes served by a parcel delivery driver was being approached as travelling salesman problem (TSP) and solved using appropriate algorithm and numerical-based optimization. The optimization approaches utilized in this problem include static routing problem, where all demand services were known ahead of time, and, delivery and pickup / collection routing did not change during execution; and dynamic routing problem, where portion of the demand services were known after some time, and, delivery and pickup / collection routing changed during execution. The formulation of a mathematical model on delivery and pickup / collection routing problem was carried out and the objective function was optimized using a two-stage linear programming (LP) technique. The first stage – static routing algorithm - was formulated and verified with another algorithm (heuristic) for correctness; static routing algorithm worked for any instance of data input. The static routing algorithm yielded an optimum total travelled distance at a minimum transportation cost. The second stage – dynamic routing algorithm - was formulated based on the previous algorithm with new challenges introduced; re-routing the static routing problem to derive an optimum routing arrangement (i.e. sequencing) at a minimum transportation cost. Both the solutions obtained were then justified with current operations for comparison in performance. Results showed that the optimal routing problem (ideal) performed 10% ~ 50% better than the current operations in term of total transportation costs and the total delivery and pickup / collection time. With an innovative use of both types of models iteratively, a decision support tool for transportation logistics was achieved.