Researchers helping to develop China’s next generation of war simulation programs are looking to nature to plot a model of conflict, with battles evolving more like organisms than machines.
In a paper published in the Chinese Journal of System Simulation this month, researchers from the PLA National Defense University’s college of joint operations said such modelling should not be an end product used in a closed setting, but constantly evolving to adapt to an open-world environment.
“The new war-game system should be treated as a ‘living organism’,” professor Hu Xiaofeng and his colleagues said.
Do you have questions about the biggest topics and trends from around the world? Get the answers with SCMP Knowledge, our new platform of curated content with explainers, FAQs, analyses and infographics brought to you by our award-winning team.
“Ecological creation is a highly complex system of engineering. There is no mature technical solution to follow. It requires continuous ‘trial and error’ in practice.”
War-game systems are used to try to predict an outcome of a conflict and have traditionally been treated as a physical process. But that approach is seen as less effective as modern military operations become more complex.
The People’s Liberation Army has focused its resources on developing war-game technology based on biological models instead, according to the researchers.
In a separate paper published in the same journal, Si Guangya, chief engineer with the PLA’s war-game program, said traditional war games assumed, for instance, that machines could not make a decision.
But with the deployment of smart weapons such as drones and unmanned submarines, artificial intelligence could launch an attack even without permission to kill, and this could change the pace of a battle completely, according to Si.
Many non-military elements including public opinion on social media, the safety of power lines and other critical infrastructure, and the economy, would also have to be factored in, Si said.
“Social domain modelling faces many new problems that are completely different from combat simulations, such as large-scale heterogeneous critical national infrastructure, and the behaviour of social groups,” Si and colleagues said in the paper.
But the complexity of such a war game would drain the available calculation resources, he added.
Hu and colleagues said they modelled war as an ecological system and nature offered some efficient solutions to the increased complexity.
For security reasons, most war games were run in a physically isolated computer network. China’s new war-game system, however, would sacrifice some security so that it could use various methods including web crawlers to refresh its information base.
The software architecture of the system also took inspiration from DNA, evolving constantly while maintaining high stability and reliability in its performance, according to Hu’s team.
The researchers said the system would need help to accomplish certain tasks in the early stages but it could improve by learning from military exercises or real operations.
“Our goal is that there will be no difference between humans and machines. They work seamlessly together to accomplish a task,” they said in the paper.
China had not fought a large-scale war for decades, so data from combat was limited. To accelerate the evolution of the system, Hu and colleagues built a “digital twin” for the system so it could learn by playing against itself.
The paper did not say when the system would be deployed officially but according to an article posted on the defence ministry’s website in October, the tentatively named “War game 2.0” system developed by Hu’s team was “ready to come out”.
There was an “explosive growth” of war-game applications in 2019, according to the article. At least one researcher had died due to work pressure. Most members in the program were “very young”, according to Hu.
More from South China Morning Post:
This article How Chinese military science is looking to nature to design war games first appeared on South China Morning Post