Comparison of RUST and C# as a Tool for Creation of a Large Agent-Based Simulation for Population Prediction of Patients with Alzheimer's Disease in EU

摘要

Introduction: During a creation of an agent based simulation is important to choose appropriate tools suitable for simulated topic. Large scale agent based simulations with a wast number of agents are very demanding on computational resources. Aim: Aim is to compare two approaches (1) Object oriented written in C# and (2) low level in Rust for creation of agent-based simulation and point out pros and cons of these tools. Methods: Simulation with over 500 millions agents is created in RUST and C#. Performance results from repeated simulation runs are compared. Runtime, memory consumption, multithreading, random number generation and agent synchronization is discussed. Results: Significant savings of computational resources were observed in Rust simulation. It's advanced options of memory management and concurrent programming options resulted in fractional execution time and memory consumption compared to C#. However, implementation in C# took less time to create and can be easily understood by domain experts.