Lightnews — Scholar-powered news

Geoffrey De Smet @geoffreydesmet.bsky.social · 2d

"Why haven't they fixed this yet?" @smarks.bsky.social and Maurice talking about ConcurrentException etc in #java at #devoxx

Great talk, but when they cover Vector and Hashtable, I can't stop and think:

"Why haven't they deprecated this yet?"

2 9

Reposted by Geoffrey De Smet

Tom Cools @tomcools.be · 3d

#Devoxx talk trial run in front of a "live" audience. 😅 If you are at the conference, do say hello when you see me. 😊

3 2 23

Geoffrey De Smet @geoffreydesmet.bsky.social · 5d

An Open Source project needs a real-time chat for the community to come together.

So we created a Discord chat for Timefold Solerv.

To ask questions.
To discuss ideas.
Or to figure out how to hack the solver to run your crazy experiment.

Join us:
discord.gg/976RcEVVHW

1

Geoffrey De Smet @geoffreydesmet.bsky.social · 20d

In the Optimization4All podcast, Cristina Radu and I talked about Operations Research, starting a company and Timefold.

Watch the podcast:
www.youtube.com/watch?v=rrLb...

Geoffrey De Smet: From a hobby to building a company

YouTube video by Optimization4All

www.youtube.com

1 1

Reposted by Geoffrey De Smet

Stephan Janssen @stephanjanssen.be · 23d

Today is the day... 🤩☕️🔥 #Java

2 13

Geoffrey De Smet @geoffreydesmet.bsky.social · Sep 6

This way he won't be to install himself in a train seat.

1 1

Geoffrey De Smet @geoffreydesmet.bsky.social · Aug 25

If it takes 10 000 hours to master a craft,
what do you get for 10 000 commits?

Timefold Solver, our Open Source solver for complex scheduling and routing problems, just reached 10 000 commits on GitHub:
github.com/TimefoldAI/t...

9

Geoffrey De Smet @geoffreydesmet.bsky.social · Aug 1

Can the optimal solution for a Traveling Salesman Problem (TSP) have crossing paths?

No, it cannot.
Because of the Triangle Inequality principle.

But in reality, it can.
Because there's road infrastructure: highways, one-way streets, etc.

In the real-world, solution optimality is rarely obvious.

1 5

Geoffrey De Smet @geoffreydesmet.bsky.social · Jul 9

A great plan is worthless if you can't explain why.

For the operators to trust your software, they need more than a button to "Schedule with AI" that generates the optimized schedule for their employees.

They need explainability. Learn how by @tomcools.be :
timefold.ai/blog/explain...

How to build trust in planning optimization: Explainable PlanningAI

Optimization algorithms can spit out mathematically brilliant schedules, but if planners can’t see the rationale, those “perfect” plans end up in the…

timefold.ai

4

Geoffrey De Smet @geoffreydesmet.bsky.social · Jun 26

Vehicle Routing Problems in production:
When research meets reality.

Our webinar yesterday covered:
- Maps integration
- Time calculation
- Hard constraints in reality
- Multi-objective optimization
- Real-time planning

See the recording:
www.youtube.com/watch?v=l12G...

Vehicle Routing Problems in production: When research meets reality

YouTube video by Timefold

www.youtube.com

1 2

Geoffrey De Smet @geoffreydesmet.bsky.social · Jun 21

That moment when your backend algorithms optimize a vehicle routing dataset with thousands of visits...

... and your frontend turns into a piece of modern art.

2

Geoffrey De Smet @geoffreydesmet.bsky.social · Jun 15

Doh. I meant float instead of int.

2

Geoffrey De Smet @geoffreydesmet.bsky.social · Jun 15

This is what happens if you use floating point numbers (int, double in #Java) for financial data.

Use decimal numbers (BigDecimal in Java) instead.

2 15 33

Geoffrey De Smet @geoffreydesmet.bsky.social · Jun 6

Where does your enterprise stand on the scheduling maturity ladder?

Are your resources scheduled manually?
With or without constraint verification?

Or automatically?
With or without optimization?

Discover the 4 Levels of Scheduling:
timefold.ai/blog/the-4-l...

Geoffrey De Smet @geoffreydesmet.bsky.social · May 18

Can LLMs solve scheduling or routing problems?

Can GenAI help companies save millions of dollars by optimizing Vehicle Routing, Shift Rostering or Job Scheduling to increase productivity and retention?

@tomcools.be and I tried it out:
www.youtube.com/watch?v=m12A...

#VRP #AI #PlanningAI #DevoxxFr

2 5

Reposted by Geoffrey De Smet

Tom Cools @tomcools.be · Apr 28

Our team has just released Declarative Shadow Variables (Preview) with our #opensource Timefold Solver 🤩 . It's easier than ever to calculate shadow variables.

As a preview feature, the API is not yet set in stone and we are actively looking for feedback. 👍

github.com/TimefoldAI/t...

Release Timefold Solver 1.22.0 · TimefoldAI/timefold-solver

We're even more excited than usual to bring you this release. Sure, it has the usual fixes and improvements, but it also brings... drumroll please... Declarative shadow variables (Preview) We've he...

github.com

1 7

Geoffrey De Smet @geoffreydesmet.bsky.social · Apr 21

Most publications of the Vehicle Routing Problem only cover the tip of the iceberg.

Benchmarks and academic papers typically only handle vehicle capacity (CVRP) and time windows (VRPTW)

But production deployments also need to deal with many other requirements.

#OperationsResearch #PlanningAI #AI

1 4

Reposted by Geoffrey De Smet

Devoxx UK @devoxx.uk · Apr 19

Discover how AI can step up when LLMs fall short in scheduling. Join @tomcools.be and dive into Timefold, a Java-based AI solver that optimises schedules using advanced math. See live demos and enhance your problem-solving skills!

Session details @ buff.ly/TePAVmY

2 12

Geoffrey De Smet @geoffreydesmet.bsky.social · Apr 10

But it doesn't stop there.

We're working on two exciting, big new features for Timefold Solver. Both are game changers to tackle real-world Operations Research problems.
More on that later. Stay tuned!

Geoffrey De Smet @geoffreydesmet.bsky.social · Apr 10

Timefold Solver is still the foundation of everything we do at Timefold, so we continue to invest in it, with our dedicated solver team.

Since we forked from OptaPlanner, we've made it faster, easier to use and more flexible to deal with real-world requirements, such fairness and resource limits.

1

Geoffrey De Smet @geoffreydesmet.bsky.social · Apr 10

Our Open Source solver now has a dedicated home: solver.timefold.ai

It's a place for the Timefold Solver community,
with everything you need to build a solution
for any planning, scheduling or routing optimization problem.

Is there anything missing that you would like to see on solver.timefold.ai?

1 1 5

Geoffrey De Smet @geoffreydesmet.bsky.social · Mar 26

What's the use case you're dealing with?

Geoffrey De Smet @geoffreydesmet.bsky.social · Mar 26

How to minimize driver costs for the Vehicle Routing Problem.

In this video, I cover dealing with a mixed fleet of employees and contractors. As well as overtime costs.

www.youtube.com/watch?v=2Foa...

#OperationsResearch #VRP #FieldService #AI #PlanningAI

Driver costs for the Vehicle Routing Problem

YouTube video by Timefold

www.youtube.com

1

Geoffrey De Smet @geoffreydesmet.bsky.social · Mar 25

The quickstarts readme for our open source solver had a make-over!

It's now easier to find the right example and start solving your scheduling or routing problem with #PlanningAI.

Try it out:
github.com/TimefoldAI/t...

3

Geoffrey De Smet @geoffreydesmet.bsky.social · Feb 11

Thanks :)