WL1
WL1
WL1
Cross-docking is a process where goods from suppliers are delivered directly to the customers,
with minimal storage and material handling through a cross-dock yard. It is a strategy widely
used in the supply chain and logistics industry. This is to improve efficiency and productivity
in the industry. In a cross-dock, minimal storage space is needed just for product consolidation
as the goods do not stay there for long. However, for traditional warehouses, large amount of
storage spaces is needed to keep the goods. They must wait until there is a customer order then
it will be picked and delivered. Hence with cross-docking, it helps to reduce the total costs and
increases customer satisfaction with a shorter delivery time. However, to achieve a good
product flow in the cross-dock system, extensive use of planning software is needed. In this
study, issues on the vehicle routing and the scheduling of the vehicles is considered. A
mathematical model is proposed with the use of a meta-heuristic algorithm, Tabu search.
In this study, the literature on the cross-dock operations is done first. It covers the problem
faced on different operating levels. It then focuses on the vehicle routing and scheduling
Followed by an extensive review on the model used with the incorporation of the Tabu search
algorithm. The use of the aspiration criteria is also included in this study to allow better and
Lastly, the results showed that despite having long computation time, an optimal route for the
deliveries can be obtained which can minizine the total cost and lead to better customer
satisfaction.
ACKNOWLEDGEMENTS
The author would like to express his heartfelt gratitude and appreciation towards the following
Firstly, Associate Professor Rajesh Piplani, for his patience, guidance and trust that he gave
the author to complete the project. Without his guidance, the author would not have the luxury
Next, PhD student Prakash Rakesh, for taking his timeout to guide the author despite his busy
schedule. His extensive help allowed the author to better understand the use of MATLAB
Finally, to the friends who have helped the author along the way when he needed advice and
1. INTRODUCTION
Cross-docking is an operation process where goods from suppliers are delivered directly to the
customers, with minimal storage and material handling through a cross-dock yard.
(Viswanathan, 2000). As speed and productivity is key to the growth of the organisations,
cross-docking is a strategy widely used in the supply chain and logistics industry. The
difference between a traditional warehouse and a cross-dock yard is the process flow in the
facility itself. For a traditional warehouse, large amount of storage space is needed to keep the
products. Only when there is a customer order then it will be picked and delivered. For the
cross-dock, goods will arrive and might not even need storage space. They will be consolidated
and being delivered. This can significantly reduce the costs that are involved in this cross-
docking process.
However, there are several constraints to achieve a smooth product and process flow in the
cross-dock. The constraints are mainly vehicle routing and scheduling problem (VRSP), the
cross-dock layout design and the network design. The network and layout design problem have
been researched on by many, but not on the VRSP. Minimizing of total cost in the cross-dock
system was done through optimization of the VRSP (Young Hae Lee, 2006). However, no
Truck scheduling in the cross-dock yards were also researched on. However, fixed outbound
schedules were not delved into (N. Boysen, 2010). Numerous algorithms have been designed
and proposed to solve the VRSP. (Wu, 2015) proposed using 4 different heuristics algorithms
to solve the VRSP, but real sized problems could not be solved. (José Brandão, 1997) proposed
incorporating Tabu Search (TS) algorithm to develop a mathematical model to solve the VRSP.
Due to some intrinsic differences between the problems, it could not be implemented on real
sized problems as well. (Agustina, 2016) proved that TS algorithm has the most potential in
solving the VRSP. Nonetheless, there is still much research to be on TS algorithm. The
previous work done by (Agustina, 2016) only covered the basis of short-term memory function.
Components such as the aspiration criteria could be considered to achieve a better solution for
This paper proposed and developed a mathematical model to solve the VRSP through the
completed. Firstly, the aspiration criteria was included into the mathematical model which
ensured that if a better solution is classified as forbidden, it would be allowed to override the
current solution. Next, to optimise the VRSP, a program was generated. Lastly, in the process
of testing the model on real sized problem the algorithm was validated.
Firstly, the constraints in the VRSP and how TS algorithm can help with the problem were
discussed in this study. Next, a mathematical model was generated using TS algorithm that was
needed to solve the VRSP. Subsequently, it showed that by using aspiration criteria, a better