BPM controls scientific quantum simulations
HQS Quantum Simulations GmbH
Initial situation
HQS Quantum Simulations GmbH, our client, offers simulation software for materials scientists in the chemical industry. Developing functional molecules or solid-state compounds requires many subsequent synthesis, formulation and testing steps to achieve desired properties. For this reason, preliminary simulations at the quantum mechanical level are the key to a deeper understanding of mechanisms and processes in molecules and materials.
Several challenges are required to provide such simulations. On the one hand, these scientific simulations can be very time-consuming and resource- and data-intensive. On the other hand, running a simulation end-to-end is a very experimental and iterative process.
Examples of this are
- handling error reports
- rerunning individual simulation steps
- further use of intermediate results
- required user interactions during the simulation
- simulation program parameterization
- using different hardware depending on the problem and simulation path
This variety of possibilities and elaborate process surrounding the actual simulation is difficult for materials scientists to handle.
This variety of possibilities and elaborate process surrounding the actual simulation is difficult for materials scientists to handle.
Approach
Novatec was commissioned to support HQS in developing a software prototype for carrying out simulation experiments. The goal was to allow materials scientists to focus on creating simulation programs and to provide easy-to-use process support for all other parts of the simulation process.
To achieve this, we modeled a process that scientists could trigger using the widely accepted Business Process Modeling and Notation 2.0 (BPMN 2.0) standard. This BPMN process calls a microservice which is responsible for retrieving the corresponding molecular structure and then triggers a simulation that is executed in a separate simulation environment. Once the simulation is complete, results are saved and made available to the scientists
Results
Camunda’s orchestration engine is used for organizing microservices and can inherently handle challenges posed by distributed environments, such as error handling and state storage for long-running processes.
Camunda itself typically supports Java-based projects. However, these simulations are performed using software packages that are implemented in different programming languages, mostly Python. We used External Task Client Pattern for this, which allows Camunda to communicate with applications developed in languages other than Java.
The molecule structure retrieving microservice was developed in Flask and communicates with a MongoDB database. The entire architecture was developed based on loosely coupled components to facilitate future migration to a quantum computing cloud environment.
The result is customized software that models molecule and material properties at the quantum level and provides scientists with the in-depth insights they need to find the optimal solution. The developed Camunda-based process and simulation module integration significantly simplifies scientists’ work. The end-to-end process is fully automated and allows researchers to focus on creating optimal simulations.
Business Process Management
As experts in process automation and individual software solutions, we bring your BPM systems in line with the latest state of the art. We know how your business processes can be digitized – and how to make them future-proof.
We also have the required skills and tools. This is how we provide you with a foundation that makes your business model scalable.
The Client – HQS Quantum Simulations GmbH
HQS provides software for materials scientists in the chemical industry and in academia.
Sophisticated quantum-level material models and their molecular properties give researchers deeper insights needed to find ideal solutions to their requirements. Founded in 2017, the company is Europe’s leading startup for complex quantum simulations and now employs more than 30 experts.