What is supply chain?

This article is ment to serve both as a broad introduction to the topic of supply chain—one of the many facets of industrial engineering—as well as a deep discussion of Professor Chen's recent research in this area.

Xin Chen, a professor of ISE at Illinois, is an expert in optimization, revenue management, and supply chain management. He was the winner of the INFORMS Revenue Management and Pricing Section Prize in 2009, and the third edition of his book The Logic of Logistics: Theory, Algorithms, and Applications for Logistics and Supply Chain Management was released in 2014. Chen recently spoke with William Gillespie, ISE marketing and communications coordinator, to talk about advances in supply chain science – a topic in which he has considerable expertise.

What is supply chain and why is it important?

Supply chain is, very succinctly, the movement of stuff from place to place. In supply chain management, you want to develop strategies to make the movement of physical products happen in an efficient way to minimize costs, satisfy your customers, and maintain a good service level.

Is it a term that is used only in manufacturing?

Not necessarily just manufacturing—it can also be applied to service. Any place where you have to deliver a physical product, you must look at the supply chain management system closely. To use the fast-food industry as an example, it’s an industry where you need to have the product ready to serve to customers. The supply chain is what supports customers knowing the product is going to be there when they want it.

When you are working on supply chain management and optimization, what sort of elements do you take into account?

There are many. One is risk. You have uncertainty and your demand is unknown. Let’s say you have disruption in your supply chain network: that’s uncertainty. The other part is the economies of scale. For example, in the manufacturing sector, you want to produce in batches. You don’t want to produce just one product at a time, because that’s more expensive. Supply chain also affects transportation, because when fulfilling customer orders, you try to avoid half-full trucks. There are other issues that are related, such as economies of scope. That’s when you want to combine different products in production and/or replenishment, but do so in an intelligent fashion.

Could that also include inventory?

Inventory is actually the key part. If you think about a supply chain network, where you are moving product from one place to another, all of those products are called inventory. If you use it, or a customer buys it, then you don’t call it inventory anymore. In most instances, you want to keep your inventory as low as possible while satisfying certain service level.

What sorts of science and math do you use in supply chain management and optimization?

Essentially, there are all kinds of optimization models. Modeling, as a measurer of uncertainty, is a very important part of the supply chain. We have to model uncertainty, so we have a determinist model and a certainty model, sometimes called a stochastic model. We use a lot of stochastic modeling to determine uncertainty. We use whatever tools are available and sometimes we develop our own tools. Some of these problems have their own distinct structures, so we develop some models and algorithms and theories to deal with these special structures.

Supply chain management is also taught at Illinois’s College of Business. How is ISE a unique curriculum in which to study supply chain?

Supply chain management has become a very important part of the engineering curriculum.  Also, of course, it’s very important in the business school. In the business school, it’s under the umbrella of operations management. There, I think it is more focused on understanding the management of supply chains at a strategic level. But in our engineering curriculum, when it comes to solving an optimization problem, we focus more on the mathematical models. The first question is, “How do you model it?” It’s a matter of getting it down to a mathematical problem. Once you model it, you develop an algorithm and solve it.

What’s the difference between upstream and downstream in supply chain?

It depends on whether something in the supply chain is close to the customer or close to the supply. Upstream means you are closer to the supply, downstream means you are closer to the customer. Knowing this helps you develop production policies. It’s a question of, “Do you want to wait when you see the demand to make production decisions,” or “Do you make production decisions before you know the demand?”

In an upstream supply chain, you make a product to stock, so you produce before you know the exact amount. And downstream you want to make to order, so you have a product made and you have it delivered. It’s something that’s also called a push-pull boundary (forecast demand is push, actual demand is pull). For example, Dell has an inventory of components, but they don’t want to assemble the components into a final product until the product is needed. So it’s a push-pull boundary they take to the assembly line. Of course, in the supply chain you have many other stages.  So, where you experience the boundary depends on the kind of business you have.

Can the science of supply chain be traced back to a particular time or place or scholar of origin?

It’s not easy to attribute the science to a particular time or place or person. It’s really just a matter of inventory management, which has been around for a while. The biggest developments have come in the 1980s and 90s. It started in the U.S., when a lot of manufacturing was moved to Asia. Because these companies no longer were just manufacturing in a single place, managing the flows of inventories in the supply chain became more important. It certainly has become a very popular academic subject and continues to be the main focal point of many companies.

Are there ethical or moral considerations you look at when you work on supply chain?

We’ve not tried to model that aspect. I know that some people have tried to model the effects of having a fake product or brand and how firms respond when their products are copied illegally. There are growing research activities in supply chain management taking into account issues such as sustainability and social responsibility.

Can you explain the term supply chain resilience?

That’s an issue that appears when you have uncertainty. Having resilience means you have the ability to adapt to the often unpredictable changes in a robust way and protect your supply chain. For example, if you had some kind of natural disaster, it could disrupt your supply chain. So you want to have resilience so your supply will be robust. If you have a good supply chain strategy, say using multiple suppliers or locating facilities in geographically separate regions, even with the disruptions, you can still maintain a healthy business.

How has network computer technology, the internet and world wide web, affected supply chain management or supply chain science in the past two decades?

The impact is huge. Walmart is a good example. They use a very successful strategy called cross-docking. Under this process, logistics are centralized to reduce inventory storage, even though orders are coming from different locations. It also can help a company better manage inventory information in its stores. To have this, you need to have good information flow, which internet technology has provided. Information is crucial to formulating large-scale optimization problems, and you have to have decision support systems to do that. It gives Walmart the ability to support a lot of trucks and to quickly update and transmit information. Another good example is Amazon.com, whose existence attributes to the development of the internet and the development and deployment of advanced supply chain management strategies.

How has e-commerce affected supply chain?

It’s had a huge impact. Over the years, we have witnessed the struggle of physical stores and phenomenal growth of online companies like Amazon.com. Customer behavior certainly has changed a lot. Because you can order online, it helps companies reduce inventory costs, but the tradeoff is increased transportation costs and response time. Developing efficient supply chain systems that can respond to customers quickly is the challenge faced by many online businesses. It’s a good example of a network structure that we can test to develop strategies.

What are some of the current problems in supply chain that your research is focused on?

My work is mainly on managing inventory. In many cases, it involves demand and pricing decisions as well. With whatever product you have and sell online, you can collect information, and now have your own inventory information. If you change the price and the customers respond, you will be making an inventory decision and a pricing decision simultaneously, and that decision should maximize the profit. You have to know the demand or distribution of your product, and how your customers are going to respond to price. Based on that, you can learn their behavior as you develop optimization strategies. Customers, especially in the retail industry, are very sensitive to prices, and you can use price as an effective strategy. Some people call this value chain management.

Does your work tend to be mostly theoretical or mathematical, or have you focused on particular industries or commodities?

My work is more on the mathematical side. My work on pricing issues is related to the retail industry. My research, though highly theoretical, is driven by real applications in most cases. I see great opportunities applying these advanced mathematical models to help firms make smart operational decisions. I’m in the process of talking to companies to find out if there’s anything interesting they’d like to do together, as far as retail research and statistics. I recently learned that people at Amazon.com are interested in my work on perishable inventory management.

Have you done supply chain work as an advisor for ISE senior engineering projects?

Yes. I have supervised several supply chain projects over the years. One company we’ve worked with produces all types of customized boxes, for which paper is the main ingredient. Since the delivery of paper orders is quite unreliable and unpredictable, the company ends up storing rolls of papers in its production facility. We helped the company reduce its inventory by 20%. We also helped another company by developing a visualization tool for its logistics systems to facilitate its truck dispatches.

Are there professional organizations you are aligned with in your field?

There are a few. There is the Institute of Operational Research and Management Science (INFORMS), the Manufacturing and Service Operations Management Society of INFORMS, the Production and Operations Management Society, and the Institute of Industrial Engineering.

Can you tell me about the third edition of your book, The Logic of Logistics, which was released in 2014?

The first edition was printed in 1997. In the second edition, we added some of the research from my doctoral thesis, which included new developments on the integration of inventory and pricing, supply chain, and risk management. The second edition had more mathematical content than the first. In the third edition, we further extended the mathematical side; for example, adding explanations of discrete convex analysis and its applications to complex inventory models. We tried to put the most recent research into the book. It’s intended to be a textbook for graduate study.

Do you teach with it?

When I teach dynamic programming courses, I use many materials from the book either to motivate or demonstrate the applications of various methodologies. I don’t use everything in the book, but I teach the foundations from the book. The book is intended as a place to find new foundational results in addition to be an up-to-date textbook. We hope to provide updated information and results that have never been published in book form.

You recently received a National Science Foundation grant to study pricing analytics, modeling theory and algorithms. Tell me about what you’re working on.

That research is basically about dynamic pricing, which is the practice of dynamically setting the selling prices of products. When you sell a product, your consumers will see your pricing trajectory and respond to it. The project is using reference price models, in which customers facing dynamic pricing form a sense of “fair” price of a product and make purchasing decisions based on that. In the end, you want to give customers a good deal to entice them to buy your product, but you also want to maximize revenue or profit. The question becomes, “How do we analyze such an operational problem and develop solutions?” I have two doctoral graduates who are working on some related topics.

I have another NSF grant looking into supply chain cooperation. The objective is to develop cost allocation mechanisms to incentivize companies to cooperate. There are reasons involving economies of scale for companies to investigate making joint orders. It can help reduce the risk, the uncertainty, and the costs. For example, in cooperative logistics, trucking companies can share their loads and reduce deadheading. Right now we are trying to find opportunities for companies to work with each other by developing effective cost or risk sharing mechanisms.

 

Related Links