7

I have a CS background, I took some optional courses in Optimization. I stumbled upon classic transport problems like TSP and VRP. Maybe I am not familiar with the domain, but I think that there aren't a lot of software apps built around those problems. Thus, I had the idea of building one. However, I am struggling with some aspects.

  1. How to choose a variant to focus on. VRP with pickup and deliveries, CVRPs, etc. I think what blocks me is that I don't know how each variant is applied in the real world, in which domains. I don't know which variants represent more opportunities and which ones are more "saturated" i.e. for which there are some sophisticated, hard-to-compete-with solutions.
  2. How to find users to talk to and how to convince them to talk to me. I don't think that building something in an ivory tower is going to succeed.
SecretAgentMan
  • 1,895
  • 2
  • 13
  • 39
SomeThoughts
  • 71
  • 2
  • "but I think that there aren't a lot of software apps built around those problems": no, there are a lot. I don't have statistics, but I wouldn't be surprised if it's the domain of OR with the most companies specialized in it. You can find some information here https://doi.org/10.1137/1.9781611973594.ch12 – fontanf Mar 06 '22 at 08:41
  • @fontanf Okay. Do you have an idea about the scheduling field? – SomeThoughts Mar 07 '22 at 19:43
  • Scheduling is a bit easier to step in. It is usually not necessary to deal with computing travel times or to write mobile apps for the users. Still, it requires good field knowledge and there are already companies on the market as well – fontanf Mar 08 '22 at 06:07
  • @fontanf okay, according to you, what are the OR problems for which it's easier to build a business? – SomeThoughts Mar 08 '22 at 10:47
  • The OR problems related to a domain for which you have good field knowledge. Starting a business in OR is not easy, the more so if you have experience (and some contacts) neither in the field, nor in OR – fontanf Mar 08 '22 at 12:45

3 Answers3

5

How to choose a variant to focus on ?

I would recommend to not choose a variant. There are so many variants that you will be better off if you build something as generic as possible. For example, the excellent ortools routing package is able to handle many constraints and options. VRPy is developed with the same philosophy.

Kuifje
  • 13,324
  • 1
  • 23
  • 56
4

There is definitely vehicle routing software in use in industry. Quite a few years ago, I consulted with a doctoral student in our supply chain management program on his dissertation proposal, which involved truck routing. He had obtained access to a commercial routing program. I never learned the name of it or its source, but it was apparently used by commercial customers.

Some companies with internal OR groups develop their own software, which may or may not be licensed to other companies. There's a nice write-up of the ORION project at UPS here.

I suspect that industry members of the Transportation Science and Logistics Society at INFORMS would have a handle on what needs are already filled and what needs represent opportunities. I'm not sure how easy it would be to network with them if you are not an INFORMS member.

prubin
  • 39,078
  • 3
  • 37
  • 104
  • Do I need to be a researcher to join? is it a paid membership? – SomeThoughts Mar 05 '22 at 18:42
  • Researcher: no. The membership is a mix of academics (faculty and students), industry people and some members of the military. Paid membership: yes. There are discounted rates for students and retirees. (I'm guessing the latter at least does not apply to you.) – prubin Mar 05 '22 at 19:58
4

I agree with Paul Rubin, there are a lot of vendors in this space.

Here a link to the 2018 Survey from OR/MS, and the vendor list. And, this is just Vendors who respond. I think there are many more vendors than in this list.

Also, there are a lot of companies who build their own routing solutions, like this one from CH Robinson (full disclosure: I worked on that one), or have consulting firms build it for them.

The space is definitely has lots of players. But, this doesn't mean there isn't room for one more.

I would have mixed feelings on what to build. You could build an entire business focused on a small niche or even by doing custom solutions for clients. Or, if you could pull it off, do something quite large scale.

Michael Watson
  • 841
  • 5
  • 11
  • Are there fewer vendors in the space of scheduling? – SomeThoughts Mar 07 '22 at 19:52
  • 1
    @SomeThoughts I don't know that space as well. What I do know about it, it is much harder than vehicle routing and prone to needing even more custom and one-off solutions. Also, since the scheduling optimization problem is so difficult, you will see many heuristic solutions in this space. I don't want to discourage you, there are many great OR problems where you can build a business- including routing and scheduling. Reach out to me if you want to chat more. I love to see people trying new things. – Michael Watson Mar 07 '22 at 21:57