Solve, simplify, and automate your
Resource Allocation, Task Scheduling, Simulations
with Decision AI aaS.
Algomia is your B2B forge for
Decision AI as a Service.
Whom is it for?
If you have a resource allocation or task scheduling challenge, we are there for you.
Juliet
My job is to manage an operational department, whose success is measured by resource-oriented KPI. I know that AI-driven decision support would be of great help. But I don't have the budget for a large upfront investment.
Cornelius
I run a SaaS company. Our solution has a scheduling module, which is essential for our users. But it is all manual. An automatization would be great, but I don't have the developers with the needed skills in my company.
Michael
Within our transformation program, we need simulations, how our changes will impact the operations of the company. I am not looking for shiny slides, but realistic variants, based on sound data to be analyzed with KPI dashboards.
10x
faster time to solution
100%
automated decision support
10+
business use-cases covered
How do we help?
We bridge the gap between your use-case and a Decision AI, which is right for you.
Do you have an ongoing painpoint with planning your resources efficiently, accurately, and at employee satisfaction? Scheduling is timeconsuming, expensive and subject to errors? Then an optimizer based on Decision AI is the right solution for you. It is a software that produces automatic, algorithmic decision support for your planners at the rules that apply to your company. We are experts in the field.
- Together, we analyze the requirements to solve your use-case.
- We configure our Decision AI models and deliver a proof-of-concept.
- After validation, the API-based solution is brought to production.
- We operate your solution, reducing your effort and improving your ROI.
Talk to Alexander
We help our clients to tackle their most intricate scheduling challenges. Our educated team of experts is ready to take it up.
Alexander Souza is the founder and CEO of Algomia. He has many years of professional experience in the optimization and software industry. He holds a doctorate degree from the ETH Zurich and holds a lectureship on algorithms at the University of Zurich.
Let's Talk!
50+
crew depots
20000+
train legs
Success Story: Large-Scale Rail Crew Scheduling
Our techology can solve even the most massive scheduling challenges.
Our customer - a top-tier national passenger rail company - wants to organize all its 20'000+ daily train legs into as few as possible shifts for its 50+ crew depots. Of course, shift consistencies, labor rules, crew qualifications, etc. must be fulfilled.
Creating a planning solution manually means a multi-month effort of a dedicated team.
1.3 bn
variables
36
threads
43h
compute time
1.9%
optimality gap
“The Algomia optimizer is seamlessly integrated with the existing planning software. The runtime and quality of results open up new opportunities on how we can approach our daily business.”
Jonas J.
Configure the Decision AI
We created a Decision AI model with 1.3+ billion variables (yes, billions, not millions). All of the required constraints are captured.
Solve the Model
Our AI algorithms solve the model within only 43h of compute time with provable optimality guarantee of 1.9%. All that happens not with specialized AI chips, but with community hardware.
Deliver the Solution
Our Decision AI algorithm reduces the overall FTE count by an undisclosed, but significant amount, translating to a substancial savings potential.
Showcase: Shiftadviser
Shiftadviser is our showcase product of an AI assistant for automatic employee scheduling.
Say goodbye to tedious and time-consuming manual shift planning. With one magical click, you quickly receive an optimized schedule.
Minimally invasive implementation
The integration of Shiftadviser into your company is minimally invasive, without deleting or affecting existing system structures.
Supplement for existing tools
Do you use Excel for your duty scheduling? No problem! Shiftadviser can be integrated using an Excel template. This way, you don't have to reinvent the wheel and can improve existing processes without replacing them.
Customized duty scheduling
With just one click, you can create a logic-based, robust and accurate schedule that takes all relevant factors into account.
“With Shiftadviser, I was able to reduce the planning time for the 50 employees in gaming from approx. 25 hours to approx. 8 hours. The optimizer automatically takes qualifications, rest periods and employee requests into account.”
Jan K.
Head of Landbased Casino, Grand Casino Baden
“The ward managers think that the duty roster is much better than before: better distributed and few gaps.”
Monique K.
Head of Staff Planning, Feldheim Reiden
“Shiftadviser's customizations for us have led to some improvements and we are increasingly satisfied with the tool's planning.”
Frank E.
Head of Table Games, Swiss Casinos Zurich
Our Decision AI Technology
Large-Scale, Mathematically Sound, API-deployed Optimization.
Resource scheduling is one of the central operational and planning challenges for many industries, e.g., logistics, transportation, health-care, entertainment, and others. We have designed and implemented a algorithmic framework, based on edge-constrained network flow, and put it into practice. It is especially well suited whenever tasks, e.g., shifts, have to be allocated to resources, e.g., staff, as to minimize some (linear) cost function, subject to respecting a mix of local and global constraints. Specific applications are Crew Diagramming and Shift Scheduling problems. Our algorithms yield very favorable running times and consistently produce high-quality solutions in practice.
- Our mathematically sound technology provides solutions with provable optimality guarantees - fast.
- See our recent abstract, which was published at the conference MAPSP - Models and Algorithms for Planning and Scheduling Problems, 2024, here.
- Scalability is incorporated algorithmically by column-generation and by cloud deployments.
- The technology can be used with a simple REST API call. Don't worry about the technical details. We take care of it.
Academic Education
Passing the Fire to the Next Generation.
We contribute to the academic community, by conference participation and by teaching algorithmic topics on a regular basis at the University of Zurich. Our prior work has led to publications in the computer science, mathematics, and health care domain. We believe that life-long learning is the basis for successful innovation.
10+
courses taught
3
theses supported
10+
internships provided
This lecture covers central and classical results in the area of combinatorial optimization. In particular, the design and analysis of "combinatorial" as well as "approximation" algorithms are treated. "Combinatorial algorithms" are exact and (mostly) polynomial-time methods, often based on dynamic programming, graphs, and linear programs. "Approximation algorithms" produce (potentially sub-optimal) feasible solutions for (usually NP-hard) computational problems. The quality of these solutions is determined by comparison against an optimal solution. The analysis will be an important and integral part. That is, we will not only state the properties of an algorithm, e.g. its correctness or running time, but also prove them mathematically. In particular, we treat the following topics: Introduction; Greedy Algorithms: Minimum Spanning Trees, Set Cover; Network Flows: Maximum Flow, Minimum Cost Flow, Assignment; Matchings: Blossom Algorithm; Linear Programming: Polyhedra, Simplex; Knapsack: Exact Algorithm, FPTAS; Bin Packing: Hardness, Heuristics, APTAS; Set Cover: Greedy, Primal-Dual, LP-Rounding; Makespan Scheduling: Identical Machines, Unrelated Machines; Satisfiability; LP-Rounding, Randimization, Derandomization.
This lecture covers the design and analysis of "randomized" as well as "online" algorithms. A "randomized algorithm" is allowed to use randomness in its decision-making. In contrast, an "online algorithm" must make decisions "on the fly" before all of the information is available, and thus must be able to decide "under uncertainty." The analysis will also be an important and integral part. That is, we will not only state the properties of an algorithm, e.g. its correctness or running time, but also prove them mathematically. In particular, we to treat the following topics: Introduction; Linearity of Expectation; Bounds on Probabilities; Markov Chains; Randomized Rounding.
Address
Algomia GmbHSchachenallee 29
5000 Aarau, Switzerland
Phone
+41 62 544 79 60
Company Registration
UID: CHE-461.339.375
VAT: CHE-461.339.375 MWST