Exploit the full Potential of your simulation Models

Most simulation models are used to characterize and locally optimize individual solutions. Based on your simulation model, Prosim‘s software solution identifies global potentials in the space of all possible solutions in the form of trade-offs of your targets. From the multitude of optimal solution alternatives that Prosim offers you can choose the solution that best suits your needs.

Efficient Data Generation

Prosim AI manages to identify trade-offs to target indicators as accurately as possible through methodical approaches and the corresponding tool chain based on the smallest possible number of simulations.

Automated Optimization

Prosim AI ties in with your existing simulation environment and triggers automated simulation runs with various system alternatives. The automated optimization will allow you to efficiently reach your goals.

Edge Technology Methods

Our methods enable a systematic approach to identifying global potentials of complex problems. They create a considerable competitive advantage over conventional approaches.

How It Works

Prosim offers an innovative methodical approach including the software that allows to implement this method. The software serves as a tool and understanding the methodological basis is key to the sucesfful use of this software. Prosim follows a generic workflow:

Customer Workshop

Before Prosim optimizes your data, we get toegether in a workshop to identify dependent design variables, their limits, as well as independent targets.

Interface Implementation

Prosim implements an interface in order to allow our optimization software to automatically communicate with its digital twin.

Optimize Solution

The optimization software uses our algortithms to determine points in the space of possible solutions.

Identify the Optimal Solution

As a result, Pareto optimal solutions are identified by Prosim according to the previously defined targets. The process characterizes the limits that can be reached in relation to the target variables using an innovative technological approach.

Result in the Personalized Dashboard

The results of the optimization of Prosim are clearly visualized in a personalized dashboard. Furthermore, the Prosim Dashboard provides a quantitative basis for communication between your product managers and development teams.

Prosim accompanies you both through the methodically substantiated workflow and through the integration of your simulation tools, up to the interpretation and presentation of your results. Together, we will create more potential from your existing modelling and simulation tools.

Successfully Implemented Projects

Optimization of Hybrid Powertrains of Trains

Due to current technologies and increased demands on emission targets, hybrid powertrains are increasingly used in trains. This raises the question of how big the batteries and the engine have to be to ensure that they work together perfectly for the train?

Optimization of Neural Networks for Bee Tracking

In order to bring the selected performance indicators of a neural network into the desired range for the application, an optimal adjustment of the hyper parameters is required. Prosim individually finds the best compositions of the hyperparameters for neural networks, efficiently and time-saving.

FAQ

Frequently Asked Questions

A software solution that identifies the Pareto-optimal solutions or trade-offs to target indicators for your defined problem based on a minimum number of simulations. We use your existing simulation models, which you already use for decision making, and connect them via an interface to our software soloution. While conventional simulations use models to characterize and locally optimize individual solutions, we use our Prosim AI to identify global potentials in the space of all possible solutions. We generate the required data for you, from which you can then select the solution that best suits you and your company.

We use neural networks that mathematically map the system relationships between input and output variables. Based on these networks, we use machine learning approaches in the parameterization of new simulation runs to efficiently identify the Pareto-optimal solutions for your problem.

You need a simulation-capable model of your system. Since our software solution then controls this model via an interface provided by us, and cannot offer any added value without such a model. Otherwise nothing stands in the way of a cooperation!

As always, a software solution is only a tool whose use requires an understanding of its methodological basis. To ensure this we follow a generic workflow:

  1. Before we optimize, we hold a workshop with you to discuss and indeitify
    1. dependent design variables (i.e. variables that can be determined by a designer) and their limits, as well as
    2. independent target values (typically set by a product manager)
  2. In the next step we implement an interface so that our AI can communicate automatically with its simulation model
  3. The AI now iteratively determines system alternatives in the design space for simulations and triggers them via the interface
  4. Im Ergebnis identifizieren wir Pareto optimale Lösungen bzw. Technologie-bestimmte Trade-offs zu den definierten Zielgrößen aus dem Workshop. Diese Trade-offs charakterisieren die Limits, die im Rahmen eines innovativen, technologischen Ansatzes in Bezug auf die Zielgrößen erreicht werden können
  5. The results of our optimization are made available to the customer in a dashboard.

We offer:

  1. A data set with all Pareto-optimal solutions
  2. Raw data of all simulations performed
  3. A dashboard for data analysis

Die Ergebnisse unserer Optimierung werden ihnen in unserem Dashboard anschaulich und intuitiv zur Verfügung gestellt. Das Dashboard stellt damit eine Art quantitative Kommunikationsbasis zwischen Produktmanager (Zielgrößen orientiert) und Entwicklungsteam (Design orientiert) dar. Zudem bietet es Unterstützung bei der Entscheidungsfindung durch interaktive Slider, die die Trade-offs greifbar machen. Für die Entwicklung gibt es ein weiteres Feature, das es erlaubt genauer die Systemzusammenhänge zwischen Ein- und Ausgangsgrößen zu verstehen.

Wir können die Datensicherung lokal bei Ihnen Firmenintern gewährleisten. Dabei unterliegt der Austausch der Daten mit unserer KI einer Datennormalisierung. Alternativ bieten wir auch die Möglichkeit die Daten bei unserem Partner in der Cloud zu speichern. Da unsere Softwarelösung an diesem Punkt flexibel ist, liegt die Entscheidung ganz bei Ihnen.

Currently we are able to efficiently identify global trade-offs in up to 30-dimensional design spaces. With Monte-Carlo or Latin Hypercube approaches this is almost impossible. In a project together with the MIT on the topic of "innovative drive systems for rail vehicles" we were able to show an efficiency advantage of over 100% in relation to non-adaptive approaches. In addition to this application, we also bring experience from a variety of other fields of application, from the design of power supply systems to complex control systems. The possibilities of our optimization are unlimited.

Have we aroused your interest?

Despite our passion for artificial intelligence, we like to maintain our contacts personally. We look forward to supporting you in the full potential development of your company. Tell us about your ideas and contact us.