Today’s post is about collaborative logistics, one field of study in which I recently started to work in. It is a tricky field for many reasons, and I will try to explain why and how. Operations research is a key enabler of collaboration. Logistics is essentially management of the activities of transporting and holding materials from origins to consumers. It has its roots in military science.

In collaborative logistics, two companies coordinate their activities in order to lower costs or improve service. Two manufacturing companies from Canada and that ship less-than-truckload (LTL) to the United States of America could pool their transportation needs so as to ship full truckloads and thus reduce total cost. In many cases, savings of 5% or more are possible by this kind of collaboration. Another key idea to reduce logistics costs is backhauling. If you ship a truckload from A to B, then the truck might return empty from B to A. If you can find another company whose need is from B to A (or close to that), you can share the cost of the full trip and save serious money.

From my example, it may seem easy, but in fact, there are some important hurdles along the way. First and foremost, it requires companies to share some sensitive information: where are your customers, what you sell them, and how much you used to pay for your transportation. You have to share that at the very beginning of the project, to get an idea of how much you would save by working together. If one of the companies later says, finally, that it’s not interested, then, it still has all the information from the other company. That’s why getting a neutral third-party to collect the data and compute these savings is good idea.

#### How OR can help?

Getting a feasible collaborative logistics plan is not simple. In particular, the question of who pays how much of what is not obvious. Many elements have to be checked out, and those familiar with game theory should have a pretty good idea of those: the plan has to generate savings for all companies, as well as making sure that no individual company is better off by itself than with the collaboration.

Although for some simple “games” – aka situations – there are some analytical formulas to determine whether collaboration can be made or not, in many cases, it’s very difficult. There are also formulas to share the benefits of collaboration among participants, but sometimes the solutions are non-practical, as they are grossly unfavorable to one of the participants. Fortunately, many of these “games” can be formulated as mathematical programming problems (often LPs). If the model is feasible, then a plan meeting all the requirements exists; if not, then we know the collaboration is unlikely to succeed, at least for this period, as some of the the participants have no rational interest in participating in the collaboration.

I’ll try to find a numerical example that puts into motion these rather vaporous concepts.