Network design is an important capability for 3PLs (Third Party Logistics providers).
A Nov 2011 Inbound Logistics article on picking a 3PL partner listed “Optimization” as #2 on the list. A big part of optimization for a 3PL is network design. That is, a 3PL needs network design capability to help their customers run a better supply chain.
As an example, Ed Melching, Director of Global Logistics for Navistar stated in an Inbound Logistics article that the reason they selected Menlo as a 3PL partner was their “global coverage, cross-network planning, and optimization capabilities.”
One such 3PL, Fidelitone, issued a press release discussion their use of network design tools (ILOG’s LogicNet Plus XE in this case):
“Based on customer order data, we were able to use ILOG to run several scenarios and analyze the geographic impact multiple locations would have on service and cost,” explained Chuck Perry, director, customer solutions for Fidelitone. “This fact-based analysis shifted the client’s assumptions and created a 15- to 20-percent reduction in cost for them. Additionally, the fact based nature of the analysis enables us to be more competitive with pricing as our confidence level in the analysis has proven true with each application of the technology…”
“…This is another great example of our continuous commitment to investing in technology and tools that lead to process improvement for our customers,” said Chuck Perry, director, customer solutions for Fidelitone. “By automating critical decisions like determining the most cost-effective solutions for managing a supplier network, we can help clients reduce costs and increase speed to market easily and more accurately.”
Clearly, having this capability is important to a 3PL.
One of the problems that you can run into when doing a large network design project is the run times of the models. Sometimes, the models you want to solve simply take much too long to solve or never solve at all.
Luckily, many network design problems do not encounter this problem. And, good modeling techniques still allow you to get the answer you need to the most complex problems.
However, you will likely still need to explain to others on the project or in the organization why you can’t guarantee a fast run-time. This question is relatively easy to answer, but our experience shows that people don’t believe the answer (we’ll share more on this in some future posts).
People don’t believe the answer because they’ve heard of cases of large models being solved quickly and they’ve models help give them answers to complex problems.
In any case, there is no guarantee that a network design model will run fast. Roughly speaking, the field of computer science breaks these types of problems into two classes: P and NP. Problems in P can have guaranteed run times that scale in a linear fashion with the size of the problem. Problems in NP do not scale linearly, but much much faster. So, a 10X increase in the size of the model may generate a 10000X increase in run time (pushing the run time into decades or more).
Network design problems fall into the class of NP problems. If network design software vendors could guarantee run-times, they would.
As a fun fact about the difficulty of these problems, there has been a standing $1 Million prize (since May of 2000) for anyone who can show that NP problems can be solved like P. So, if you can guarantee the run times of these problems, you can collect a $1 Million prize.
Some of the authors will be at CSCMP in Atlanta starting on Sunday Sept 30th. We’ll be at the IBM booth. If you are coming to CSCMP, let us know (firstname.lastname@example.org) and we’ll be sure to set up a meeting.
On Sept 21, on the Ops Rules Blog, David Simchi-Levi responded to a question about whether to make network design decisions with a spreadsheet or commercial tool.
He says the main reason is that Excel is not set up to handle the complexity of the problem and the mathematical optimization. In addition, he cites several advantages of a commercial package:
- Ease of use of setting up the data and validating it
- Built in optimization and analytics which cannot be done with a spreadsheet
- Built in maps, graphs, scenario comparison and reporting capabilities
- Tested and stable platform for continued analysis – easy to transfer and share with other users
If you want to see a 15-minute video of commercial tool, here is a link.
I would add just one more point: There is a lot at stake with network design. You can be making decisions that impact significantly impact your costs and service. You should make sure you understand, at least at a high level, the math involved in these before you do the project in Excel. This way, you will have a sense of what you are giving up.
On Sept 19, the Chicago Tribune ran a story from McClatchy Newspapers on companies moving production from China to Mexico as the wage gap closes.
Interestingly, besides just the cost of wages, the article sites examples of some of the hidden fixed costs in production. For example, problems will always surface in a plant. How much does it cost in terms of time, wear and tear, and money to get the right engineers to the plant to fix the problem. The article quoted one firm saying that if there there is a problem in Mexico, the US engineers can leave in the morning and start working on the problem several hours later. China requires much more.
This is similar to an article from the Wall Street Journal this summer where David Simchi-Levi’s study was quoted as showing that companies were moving production back to the U.S. He called this “reshoring.”
Among the main reasons cited for reshoring: a desire to get products to market faster and respond rapidly to customer orders; savings from reduced transportation and warehousing; improved quality and protection of intellectual property.
In a very short time, the factors that determine where you should produce can change. This also impacts your distribution network. It is not clear what the landscape will look like three years from now, but these examples highlight the importance of network modelling. A model of your business and the key drivers allows you to keep your supply chain as efficient as possible.
While writing about supply chain network design, we realized that it touched many different subject areas. We wrote the network design book so that you could use it in many different types of classes.
For example, here are nine different areas you could explore in more depth after an introduction to the topic through network design:
- Optimization. When you solve a network design problem, the solver uses mixed integer programming (MIP). The book provides the formulation to help the reader gain intuition from the math (click here for more on the value of understanding the math). If you want to your class to focus on optimization, you can start with the formulations, and then have your students explore more about MIPs. We have discussions on the knapsack problem and computational complexity you could cover in more depth.
- Transportation Markets. Obviously, transportation costs are important to network design and facility location. We introduce different modes of transportation like Full Truckload, Less-Than-Truckload (LTL), Parcel, Rail, and others. You could use this as a staring point to discuss these markets in detail.
- Modeling. We discuss how to map the real world into a mathematical model. This is a skill that should be applied beyond network design. Also relevant to other problems is a discussion about the level of precision and number of significant digits in a model.
- Activity Based Costing vs Standard Accounting. When building a network design model, you need to carefully use cost information that will help answer the appropriate question. Often, this cost information is not the same that gets reported by the accounting team. The costs reported by the accounting team may have mixed fixed and variable costs together and may have assumptions about the throughput of a site. It is important to understand these assumptions before using a cost figure in a model.
- Regression. Regression can help you determine transportation costs for lanes you aren’t currently based on the lanes you are using.
- Vehicle Routing. When you open new warehouses, your delivery routes will change. You can use network design to introduce vehicle routing.
- Inventory Optimization. Service level is an important concept in network design. But, the location of facilities only covers some aspects of service level. Inventory optimization covers other aspects. Also, as you add or remove warehouses, you may have more or less system-wide safety stock.
- Debugging. Network design models capture the entire supply chain. This means you can build some complex model. Debugging is an important topic and it applies to much more than just network design.
- Multi-Objective Optimization. Many other problems besides network design balance multiple objectives. You can use network design to introduce the topic and then go into other problems.
Our course material can help you get started with each of these areas.
var _gaq = _gaq || ;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script'); s.parentNode.insertBefore(ga, s);
James E. Cooke ran an article on Sept 18 in DC Velocity, “A Network Design is Never Done.”
This is a nice article that highlights a trend we are seeing as well. Their article quotes a Tompkins Supply Chain Consortium study that shows that the average time between studies has decreased from 24 months to 18 months. Since this is the average, many firms are likely doing studies every 6 or 12 months with some doing studies on an on-going basis.
The article sites changing customer demand, the need for contingency plans, and volatile oil prices as some of the reasons for on-going network design.
This article is another reminder of the importance of network design on the overall performance of your firm.
I haven’t thought much about pizza delivery in the last 10 years. However, it seems like a lot of people do. And, efficient pizza delivery makes people happy.
So, maybe we should use pizza deliver to explain supply chain concepts to people who may not have thought much about the supply chain. When we wrote the book, pizza delivery helped explain some concepts. Within the last 48, I’ve seen two other examples with pizza delivery.
Here are my top 3 reasons why you should use pizza delivery to explain the supply chain:
- Our book has a good example that uses pizza delivery to define different uses of the term service level. Service level can mean the time to get a pizza to someone’s house or the percent of time you have the ingredients your customers ask for. You need good network design to locate your restaurants and good inventory control and forecasting to make sure you have what your customers will order.
- Yesterday’s blog post brought up the fact that you better have good processes in place to quickly get the order, make the pizza, and deliver it. Specifically, yesterday’s post highlighted that you better be able to take orders from the phone, the web, and from mobile apps. A good discussion on having the right IT strategy.
- The book, The Human Face of Big Data, used pizza delivery in Manhattan to show movement. The picture at the top of the post shows the path’s of the Domino’s pizza delivery bikers on a Friday night. From the picture, the very bright spots show the locations of the restaurants and you can see the areas with the most demand. This highlights how a supply chain must deliver to its customers. The article points out that these restaurants receive product from a warehouse in Connecticut (so, there is an infrastructure behind the local infrastructure). Someone had to decide on this warehouse in Connecticut and how many cities it would serve.
With these three examples, you can explain a lot about supply chain strategy with a business that people can relate to.
Service level is an important concept in network design.
However, the term “service level” has many different meanings within a supply chain.
So, while writing this book, we needed a good example to help explain the different definitions of service level and how they do or don’t impact network design decisions.
One of the authors (who we won’t name) came up with the great idea to use pizza delivery as an example. She thought that after a long week at work, being able to quickly get a pizza delivered to your home can bring pure happiness. Everyone can relate to this example.
This example, which we expand on in the book, is easy to explain and describes the different types of service levels and how network design can help. For example, network design can help locate your restaurants close to customers, but it can’t help you quickly take an order and bake a pizza.
Apparently, others are also using pizza delivery to discuss service level. This blog post by Clay Richardson uses this example:
…I needed to order a pizza for my family after a very hectic Saturday afternoon. When I picked up my mobile phone to call the pizza delivery place, a light bulb went off over my head. Instead of dialing the pizza delivery company and waiting on hold for 15 minutes, why not download its mobile app in two minutes and order my pizza within another two minutes. I figured I could shave off ten minutes of wait time by simply downloading the pizza delivery company’s mobile app.
Take a guess how long it took me to order the pizza? A whopping 20 minutes! Why? Because the pizza delivery company did not invest time optimizing its order process for the mobile experience. “
This example again proves the difference between operational service level measures and those related to network design problems. Similar to our example in the book, in this situation the restaurant was located where the customer had already proven the pizza could be delivered in a time that was acceptable but the service level related to the order process causes an unsatisfied customer in the end. Both types of service are important to the delivering a pizza on time. And what else provides more satisfaction than a hot pizza delivered quickly to the customer’s while enjoying their weekend!
The expanded Panama Canal, which will allow much larger ships to pass through and will double the capacity, is expected to be completed in 2014.
The first impact is that port cities east of the canal are preparing the handle the larger ships.
The second impact is that firms will have to re-evaluate their supply chain infrastructure.
Many firms currently bring product into the US through the West Coast ports of Los Angeles and Long Beach. Many of the containers that arrive on the West Coast are shipped via rail or truck to destinations east of the Mississippi River.
So, now that larger ships and more volume can pass through the Panama Canal, does it make sense to shift volume to east coast ports?
The answer will vary for every firm, but the drivers will be:
- Cost to ship product through the Panama Canal versus the cost to ship through a west coast port
- Time to ship through the canal versus through a west coast port
- Location of customer demand
- Location of existing production and distribution facilities
I suspect that these new options will increase the need for network design analysis. Not only will the panama canal create a new option, but the west coast ports, inland ports, and railroads will likely work to become more competitive as well.