Complicated decisions

-

simply automated

As an expert for mathematical models and algorithms, I help you make intelligent decisions at the push of a button. I’m a freelancer for optimization and software development projects.
A peek into my Tech-Stack:
C++
Gitlab
Gurobi
GAMS

Analytics Report

59%

Lorem ipsum dolor
01.02.2022
01.02.2022

My services
for your success

I believe humans should be able to focus on creating what no one else could create. I empower deep focus by letting computers make decisions that involve numbers and abstract concepts, practically innumerable alternatives, and require a systematic approach. This way, computers become your most patient and enduring support in choosing meaningful actions. Here are my specific offers for you:

Modeling

I create a mathematical model of your optimization problem to describe a complicated real-world decision task in terms that are accessible to the computer. Such a model comprises decision variables, constraints, and an objective function.
Wheth­er max­i­miz­ing ef­fi­cien­cy and pro­fit, or min­i­miz­ing en­er­gy de­mand or cost - I cre­ate a math­e­mat­i­cal mod­el to de­scribe your com­pli­cat­ed re­al-​world de­ci­sion task in terms that are ac­ces­si­ble to the com­put­er. Such a mod­el com­pris­es de­ci­sion var­i­a­bles, con­straints and an ob­jec­tive func­tion.

Algorithms and Solvers

Supported by a bag of mathematical tricks, I develop and test various promising model formulations to identify the most efficient one. Efficient formulations minimize the computational cost of the optimization and maximize the utility of the resulting solution.
Sup­port­ed by a bag of math­e­mat­i­cal tricks, I de­vel­op and test var­i­ous promis­ing mod­el for­mu­la­tions to i­den­ti­fy the most ef­fi­cient one. Ef­fi­cient for­mu­la­tions min­i­mize the com­pu­ta­tion­al cost of the op­ti­mi­za­tion and max­i­mize the u­til­i­ty of the re­sult­ing so­lu­tion.

Integration

This phase is all about your workflows, respectively those of your customers. I make sure to maximize the impact of the optimal solution and bring visually prepared results, recommendations, and alternatives to the screens of your employees and customers. Additionally, the hard numbers are made available in ERP systems, Excel reports and databases.
This phase is all a­bout your work­flows, re­spec­tive­ly those of your cus­tom­ers. I make sure to max­i­mize the im­pact of the op­ti­mal so­lu­tion and bring vi­su­al­ly pre­pared re­sults, rec­om­men­da­tions, and al­ter­na­tives to the screens of your em­ploy­ees and cus­tom­ers. Ad­di­tion­al­ly, the hard num­bers are made avail­able in ERP sys­tems, Ex­cel re­ports and da­ta­bas­es.

Software Development and Consulting

In every project, you will receive tailor-made software and tools to streamline your decision and planning processes, as well as vivid explanations for all questions coming up. A well-prepared documentation of our results completes my deliverables for you.
In eve­ry pro­ject, you will re­ceive tai­lor-​made soft­ware and tools to stream­line your de­ci­sion and plan­ning pro­cess­es, as well as viv­id ex­pla­na­tions for all ques­tions com­ing up. A well-​pre­pared doc­u­men­ta­tion of our re­sults com­pletes my de­liv­er­a­bles for you.

Modeling

I create a mathematical model of your optimization problem to describe a complicated real-world decision task in terms that are accessible to the computer. Such a model comprises decision variables, constraints, and an objective function. The decision variables describe the nature of each partial decision; for example, yes/no, a number of entire units (such as trucks used), and the amount of some continuous good (such as electrical energy).
Wheth­er max­i­miz­ing ef­fi­cien­cy and pro­fit, or min­i­miz­ing en­er­gy de­mand or cost - I cre­ate a math­e­mat­i­cal mod­el to de­scribe your com­pli­cat­ed re­al-​world de­ci­sion task in terms that are ac­ces­si­ble to the com­put­er. Such a mod­el com­pris­es de­ci­sion var­i­a­bles, con­straints and an ob­jec­tive func­tion. The de­ci­sion var­i­a­bles de­scribe the na­ture of each par­tial de­ci­sion; for ex­am­ple, yes/no, a num­ber of en­tire u­nits (such as trucks used), and the a­mount of some con­tin­u­ous good (such as elec­tri­cal en­er­gy).
Con­straints are (e.g., ther­mo­dy­nam­ic) link­ages of the de­ci­sion var­i­a­bles, as well as sit­u­a­tion-​spe­cif­ic log­i­cal, le­gal, and stra­te­gic re­stric­tions on the over­all de­ci­sion. For ex­am­ple, a well-​trained em­ploy­ee can op­er­ate three dif­fer­ent ma­chines, but at most one at a time. To­geth­er, we de­scribe your qual­i­ty and safe­ty re­quire­ments, col­lect da­ta and pro­cess flows, as well as id­i­o­syn­cra­sies of po­ten­ti­al­ly aged plant sys­tems that have been stead­i­ly ex­pand­ed. The ob­jec­tive func­tion is a for­mu­la that as­sess­es the qual­i­ty of a de­ci­sion. As an ex­am­ple, pro­duc­tion costs of pro­cess­es with time var­i­a­ble costs can be used to as­sess a pro­duc­tion plan.
Constraints are (e.g., thermodynamic) linkages of the decision variables, as well as situation-specific logical, legal, and strategic restrictions on the overall decision. For example, a well-trained employee can operate three different machines, but at most one at a time. Together, we describe your quality and safety requirements, collect data and process flows, as well as idiosyncrasies of potentially aged plant systems that have been steadily expanded. The objective function is a formula that assesses the quality of a decision. As an example, production costs of processes with time variable costs can be used to assess a production plan.

Algorithms and Solvers

Supported by a bag of mathematical tricks, I develop and test various promising model formulations to identify the most efficient one. Efficient formulations minimize the computational cost of the optimization and maximize the utility of the resulting solution. Different modeling methods can describe exactly the same problem. These methods can lead to continuous or mixed-integer models and to strongly structured, or weakly structured models. Also, they utilize different numbers of variables and constraints.
Sup­port­ed by a bag of math­e­mat­i­cal tricks, I de­vel­op and test var­i­ous promis­ing mod­el for­mu­la­tions to i­den­ti­fy the most ef­fi­cient one. Ef­fi­cient for­mu­la­tions min­i­mize the com­pu­ta­tion­al cost of the op­ti­mi­za­tion and max­i­mize the u­til­i­ty of the re­sult­ing so­lu­tion. Dif­fer­ent mod­el­ing meth­ods can de­scribe ex­act­ly the same prob­lem. These meth­ods can lead to con­tin­u­ous or mixed-​in­te­ger mod­els and to strong­ly struc­tured, or weak­ly struc­tured mod­els. Al­so, they u­ti­lize dif­fer­ent num­bers of var­i­a­bles and con­straints.
An­oth­er im­por­tant mod­el prop­er­ty is wheth­er on­ly lin­ear terms are used, or al­so quad­rat­ic terms, or even more strong­ly non­lin­ear terms. This prop­er­ty in­flu­en­ces which al­go­rithms are suit­a­ble to com­pute a so­lu­tion for the prob­lem. In ad­di­tion to low com­pu­ta­tion­al cost, ef­fi­cient mod­els ex­hib­it the fol­low­ing at­tri­butes: nu­mer­i­cal sta­bil­i­ty, in­ter­pre­ta­bil­i­ty of the so­lu­tion, ex­ten­si­bil­i­ty and mod­u­lar­i­ty, and good read­a­bil­i­ty.
Another important model property is whether only linear terms are used, or also quadratic terms, or even more strongly nonlinear terms. This property influences which algorithms are suitable to compute a solution for the problem. In addition to low computational cost, efficient models exhibit the following attributes: numerical stability, interpretability of the solution, extensibility and modularity, and good readability.

Integration

This phase is all about your workflows, respectively those of your customers. I make sure to maximize the impact of the optimal solution and bring visually prepared results, recommendations, and alternatives to the screens of your employees and customers. Additionally, the hard numbers are made available in ERP systems, Excel reports and databases.
This phase is all a­bout your work­flows, re­spec­tive­ly those of your cus­tom­ers. I make sure to max­i­mize the im­pact of the op­ti­mal so­lu­tion and bring vi­su­al­ly pre­pared re­sults, rec­om­men­da­tions, and al­ter­na­tives to the screens of your em­ploy­ees and cus­tom­ers. Ad­di­tion­al­ly, the hard num­bers are made avail­able in ERP sys­tems, Ex­cel re­ports and da­ta­bas­es.
In­ter­ac­tive­ly with the re­spon­si­ble u­sers, I es­tab­lish a way to cap­ture im­por­tant da­ta, such as cur­rent in­ven­to­ry lev­els or the or­der si­t­u­a­tion. Thus, eve­ry prob­lem in­stance that is solved is based on up-​to-​date in­for­ma­tion. Tai­lor-​made in­ter­pre­ta­tion aids for the so­lu­tion en­sure that us­ers will be look­ing for­ward to re­ceiv­ing the re­sults. Ide­al­ly, a suc­cess­ful in­te­gra­tion phase al­lows the prob­lem to be spec­i­fied, solved and in­ter­pret­ed at the push of a but­ton.
Interactively with the responsible users, I establish a way to capture important data, such as current inventory levels or the order situation. Thus, every problem instance that is solved is based on up-to-date information. Tailor-made interpretation aids for the solution ensure that users will be looking forward to receiving the results. Ideally, a successful integration phase allows the problem to be specified, solved and interpreted at the push of a button.

Software Development and Consulting

In every project, you will receive tailor-made software and tools to streamline your decision and planning processes, as well as vivid explanations for all questions coming up. A well-prepared documentation of our results completes my deliverables for you. I have several years of experience developing in modern C++ standards and using the STL. Furthermore, I learned to code in older standards and am well-versed in dealing with legacy software.
In eve­ry pro­ject, you will re­ceive tai­lor-​made soft­ware and tools to stream­line your de­ci­sion and plan­ning pro­cess­es, as well as viv­id ex­pla­na­tions for all ques­tions com­ing up. A well-​pre­pared doc­u­men­ta­tion of our re­sults com­pletes my de­liv­er­a­bles for you. I have sev­er­al years of ex­pe­ri­ence de­vel­op­ing in mod­ern C++ stan­dards and us­ing the STL. Fur­ther­more, I learned to code in old­er stan­dards and am well-​versed in deal­ing with le­ga­cy soft­ware.
I al­so de­vel­op so­lu­tions in GAMS, MAT­LAB, and Mo­de­li­ca, as well as help­ful scripts. As part of your team, I feel re­spon­si­ble for ef­fi­cient com­mu­ni­ca­tion and over­all pro­ject pro­duc­tiv­i­ty in eve­ry soft­ware de­vel­op­ment pro­ject. In my day-​to-​day work, I ap­ply i­deas of ag­ile work­ing. I have been work­ing suc­cess­ful­ly in glob­al­ly dis­trib­ut­ed teams for many years.
I also develop solutions in GAMS, MATLAB, and Modelica, as well as helpful scripts. As part of your team, I feel responsible for efficient communication and overall project productivity in every software development project. In my day-to-day work, I apply ideas of agile working. I have been working successfully in globally distributed teams for many years.
Example Application Fields
1. Cost minimization
2. Production planning
3. Transportation
4. Energy efficiency
5. Energy networks or logistic network
6. Location expansion

Our Steps to Success

Applying agile working methods, I ensure efficient communication and involve key people on site. If something new is discovered over the course of the project, I can react quickly with a focus towards the project goal.

01

First Contact

5 minutes
On this website, you can briefly describe your project, idea, or concern in the contact form. Thereby, you help me optimally prepare a complimentary initial meeting.

02

Initial Meeting

1 hour
I will invite you to a confidential video-conference meeting to explore whether a mutual project makes sense for us. In the best case, we already create a first roadmap of how your requirements can be implemented until the project is completed. Afterwards, you can review my offer in depth.

03

Kickoff

1-3 weeks
Together, we agree on the first milestone, still ignoring many details. On the way to the first milestone, we get to know each other, and synchronize our communication and working styles. The kickoff phase concludes with a mutual review of the effectiveness of our roadmap for the desired project goals. We may adjust our trajectory based on early learnings.

04

Implementation

2 weeks - several months
With each milestone, we implement more and more details and test the results over the course of small iterations. The modeling phase is the heart of projects that start 'from scratch'. Even if the focus is in other areas, individual modeling details are often still adjusted.

05

Finalization

1 day
Subject to the approval of your customers and employees, we complete the project. Our user-centric perspective ensures that the project results are integrated into existing processes in a way that adds value. I will also hand over the project documentation, which serves as an effective introduction for new users and facilitates later additions.

About Tim

Hi, I'm Tim Varelmann. To give people capacity to bring their very own talents into the world, I harness computers to automate complicated, abstract and repetitive tasks. This takes care of nitty-gritty details and allows to see the bigger picture in day-to-day life.

My journey has taken me to four top universities (RWTH Aachen, University of Queensland, MIT, UT Austin) on three continents and made me an expert in mathematical optimization. In applications ranging from power grids to process engineering, I built a proven track record as a problem solver, and I am curious to go beyond these fields. Besides mathematical optimization, I enjoy playing soccer, board games, and all sorts of water sports - except swimming.

References

Don't just take my word - see what my clients have to say.

Professional and extremely competent

During the excellent collaboration with Tim, I have found him to be a very pleasant, highly productive and extremely competent person. Tim was involved in an advanced-stage development project with us and his expertise in numerical optimization was a driving force in our development process. With his know-how, he solved crucial performance problems and thus enabled us to scale our models.

Stefan Kuhn

Development Engineer at thyssenkrupp Industrial Solutions AG

Contact me

How can I help you? Which of your applications and processes involve intricate decisions and plans? How does your workflow look like? Tell me about your ideas and perspectives.
Thank you! We'll be in touch soon.
Oops! Something went wrong while submitting the form.
Do you prefer to:
Send email

My Blog

The merits of mathematical models
Model-based digital transformation is THE vehicle to move industries to the digital future. This approach is actionable from the start, fosters agile work, and delivers valuable results.
23.11.2022
How to succeed on the pursuit of a PhD
During the summer break, my PhD certificate arrived, testifying the success of my research in the development of mathematical optimization algorithms and software. The occasion made me reflect on the journey I have had and what I would do differently - this is the result.
21.9.2022
All blog posts

Start improving your decisions today!

Unleash the power of modern software and mathematical precision for your business.
Start your project now