Naslov
System Dynamics Modelling with Kotlin: From Hierarchical Models to Interactive Simulators
(System Dynamics Modelling with Kotlin: From Hierarchical Models to Interactive Simulators)

Autori
Sovilj, Siniša ; Etinger, Darko ; Sirotić, Zlatko ; Pripužić, Krešimir

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, prošireni sažetak, znanstveni

Izvornik
Proceedings of the 38th International Conference of the System Dynamics Society 2020 / - , 2020

Skup
38th International Conference of the System Dynamics Society

Mjesto i datum
Bergen, Norveška, 19.07.2020. - 23.07.2020

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
system dynamics, kotlin, modelling, complex system, hierarchical model, simulator

Sažetak
System dynamics (SD) modelling is traditionally done by graphically using mostly proprietary software or with couple of open and free software exceptions. Building large hierarchical models with reusable SD modules is not supported well and building fast and customizable simulators for interactive learning on multiple target platforms is not trivial. There are couple of framework/toolkit solutions that fulfill that gap using general-purpose languages (Java, Python, JavaScript) but they have their own limitations. We asked whether Kotlin, as a new, modern, statically-typed, null- safe, object-oriented and functional language can do any better and overcome limitations of other programming languages and frameworks/toolkits. Therefore, we started to develop Kotlin SD Toolkit which does not exist so far at our best knowledge. We found that Kotlin as a new, programming language and our Kotlin SD Toolkit as a new tool are both suitable for modelling larger, hierarchical SD models that support modules (see Table 1) and for easier development of interactive simulators for multiple target platforms: desktop, web or mobile (see Figures 1, 2 and 3). We also measured speed properties without any a priori code optimizations and for now Kotlin SD Toolkit on our computer needs 7 seconds to numerically integrate simple testing SD model with 1E7 time steps, which is very fast in comparison with others and can be additionally improved. We will continue to further expand capabilities of the Kotlin SD Toolkit as an open source project and our contribution to SD community.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo, Interdisciplinarne tehničke znanosti, Ekonomija, Informacijske i komunikacijske znanosti

POVEZANOST RADA