Square Stage

“The purpose of teaching is not to convey information but to excite and inspire”

- Teo Chung Piaw -

A little flexibility can go a long way in improving the performance of the operational system

My early work in this field explores this theme, and develop a mathematical approach to analyze problems of these types. We use the properties of graph expanders to argue that a sparse network, with good expansion properties, can perform as well as a fully flexibile system in matching supplies with demands. Early work with these results are:



 Design for Process Flexibility: Efficiency of the Long Chains and Sparse Structure,

Mabel Chou, GBA Chua, CP Teo and H Zheng. __Operations Research__, Volume 58 (1), pp 43-58, 2010.



Process Flexibility Revisited – The Graph Expander and Its Applications, 

Mabel Chou, GBA Chua, CP Teo and Huan Zheng. __Operations Research__. Sep/Oct 2011, Vol 59 no. 5 1090-1105



On Range and Response: Dimensions of Process Flexibility,

Mabel Chou, G. Chua and CP Teo. __European Journal of Operational Research__. 207(2), 711-724, 2010.



Process Flexibility: Design, Evaluation and Applications, 

Mabel Chou, CP Teo and Huan Zheng.. __Flexible Services and Manufacturing Journal__. (Invited Review), 59-94, 2008. Formerly published as International Journal of Flexible Manufacturing Systems.

How to design sparse and efficient network structure to support supply chain operations?

In recent years, I begin to investigate the above question, uaing the tool of conic programs, which has seen a lot of development in recent years. Basically, by employing a copositive cone of the underlying supply chain problems, and analyzing the dual prices of an associated relaxation, we can design very good sparse structure for the operations. This brings our understanding of flexible operations beyond the notion of 2-chains, into more general supply chain systems.


On the Design of Sparse but Efficient Structures in Operations,

 Yan Zhenzhen and Gao Yini. Forthcoming, __Management Science__.


Reliable Flexibility Design of Supply Chain via Extended Probabilistic Expanders,

 Y. Liang, H. Shen and Max Shen. Forthcoming, __Productions and Operations Management__.


On Range and Response: Dimensions of Process Flexibility,

Mabel Chou, G. Chua and CP Teo. __European Journal of Operational Research__. 207(2), 711-724, 2010.

The notion of flexibility can be extended to study the following question: How does one design the capacity of a supply chain system, if the final allocation of capacity to demand streams are decided based on real time/actual demand realized? In a nutshell, this is a generalization of the classical newsvendor problem to the multi-plant multi-item environment.

What are the benefits of real time allocation flexibility to a supply chain?


Resource Pooling and Allocation Policies to Deliver Differentiated Service,

Zhong Yuannguan, Zheng Zhichao and Mabel Chou. Forthcoming, __Management Science__

This paper shows that even in the one-plant multi-item setting, the effect can be enormous - in many large scale system, real time allocation flexibility can eliminate the needs to hold safety stock, even if the service levels demanded by the customers can be extremely high (close to 100%). Interestingly, the analysis requires the use of Blackwell's approachability theorem, which is a well knwon tool in minimax game theory.


Capacity allocation in flexible production networks: Theory and applications

G. Lyu, W. C. Cheung, M. C. Chou, Z. Zheng , and Y. Zhong. Forthcoming __Management Science__

This paper extends the analysis of the previous paper to the multi-plant multi-item setting. This brings a new perspective to the well known 2-chain phenomenon - not only that 2-chain works well in a balanced system to maximize utilization of capacities, in a service level driven setting, 2-chain can also help to minimize the capacity configurations required to meet certain service level commitments.

Interestingly, this is also where we begin to realize that the problem can be analyzed using the online convex optimization framework, a recent popular tool in the computational learning domain. Another paper that explore similar theme, but from the type-one service perspective:


Capacity allocation with stock-out probability targets: Theory and applications,

G. Lyu, M. C. Chou, C. P. Teo, and Y. Zhong.



Proceeding of 2010 8th International Conference on Supply Chain Management and Information Systems: Logistics Systems and        Engineering