AIMMS works with a global ecosystem of consulting companies and resellers to deliver powerful optimization solutions to customers. For the next installment in our Partner Spotlight series, we spoke to Jacob Jan Paulus, Senior Consultant at CQM.
CQM is a data science consulting firm based in Eindhoven, the Netherlands. They offer optimization, statistics and AI applications with focus ranging from industry, retail, and service.
CQM has been a partner almost since the beginning of AIMMS – about 30 years. How has the relationship developed over the years?
We’ve currently got about 15 specialists who use AIMMS regularly. In the last few years we’ve had some bigger projects, which brought us in closer contact with the AIMMS support team, and resulted in a growing relationship with R&D. In working together, we have shared insights about users’ needs and provided a lot of input for the development of new features.
How has AIMMS been a part of your career with CQM?
I used AIMMS while I was teaching a university class in OR, working with AIMMS founder Johannes Bisschop, so I was already very familiar with it when I started at CQM 10 years ago. At CQM we have a community of optimization specialists helping each other and sharing ideas on AIMMS based projects. For an optimization project, we always consider AIMMS as a technology.
What we really like about AIMMS projects is how quickly we can get a solution to a client. Using a more general programming language can take a long time to build up a model from scratch, but AIMMS developer tools make it quick to get the solution up and running and deployed.
What is an AIMMS-based project CQM is especially proud of?
We worked on a very successful project for AgroEnergy called Optimal Bid (BiedOptimaal) that won the Dutch Data Science Award (Nederlandse Data Science Prijzen). The problem was about forecasting energy needs for greenhouses. We had to consider weather conditions, specific needs of the plants inside a greenhouse, and fluctuating costs for different energy sources. The optimization solution we created enabled the client to choose the most efficient energy source for a given period.
You’ve seen a lot of AIMMS projects through for different business scenarios. Does one project come to mind that has had an interesting journey?
We work with FUJIFILM to make operations planning and scheduling decision apps. We have a solution to optimize their production of offset printing plates of all different sizes. Our solution minimizes material waste and production cost. It’s very important to their business, since raw materials is one of their major costs.
In this kind of solution, the client takes the results immediately into production. You can’t just give general suggestions; it must be specific recommendations based on detailed data. It’s a challenge not only to apply complex optimization models to get the most value, but also to ensure that your solution becomes a major part of their planning process to make important decisions.
What are the biggest challenges when handing off or implementing solutions?
Clients often have specific processes in place for decision making, so when you give them a new solution to support decision making, not only do they have to learn a new concept and how to use a new tool, but also accept that the tool may give them surprising outcomes that imply changes to their way of making decisions. It’s important to have good communication between our specialists and people on the client’s side who facilitate the project and lead the implementation of those changes.
With an optimization-based solution, we don’t aim to automate decision making itself, but rather change the way decisions are made. We give people the tools to make better decisions. Instead of spending time on creating just one solution, they get more insights and can then spend time evaluating multiple scenarios and aligning with their stakeholders.
What advice would you give an OR consultant just starting an AIMMS-based project?
You need to start with a high-level overview of the end users’ process. Talk to people who really know the business problem and how it relates to the business as a whole.
Of course, you need advanced algorithms and the “smart” stuff operating in the background to do the optimization, but you need to tailor it to the client. You must really understand the problem, know the limitations, and know where this decision fits into their bigger business decisions. Knowing all this before you start building a solution, you can make informed and intelligent trade-offs in the process.
You also need to think long-term. How will you maintain and make changes to the app down the road? One way to make that easier is to break a complex problem into smaller apps, with each having a solution to address a specific part of the problem.
Another important thing to consider is structure and conventions in the way you build a project. When you always structure your projects in the same way, it makes it much easier to explain and pass down to another developer later.
Connect with Jacob Jan on LinkedIn to learn more about CQM, or visit their Contacts page to get in touch.