I am looking to enter the field of Complex Systems or Complexity Science as a postgraduate in my Computer Science faculty, specifically focusing on Agent-Based Modeling. However, I am unclear on the current problems being addressed, what the research trends are, and how it may be applied in industry from a CS perspective, because there is currently no one in my faculty involved in Complex Systems. Any clarification on this issue would be greatly appreciated, thank you.
It is always good to start by wider reviews of the field you are asking for and to know all the methods used there. You will due to this broader overview get a better sense of advantages and disadvantages of your chosen method to solve a given problem!
The highest truth is that applications are shaping the methods developed and not vice verse. Hence, two things are important: (a) which area of research is developing faster in applications of CSs, (b) what is the broadest range of CSs applications applied there.
Those two questions will lead you towards desired answers. Due to my current research focusses, I will dive into medicine and CSs used there.
There is a fast-growing field of complex systems applications in biomedical research and medicine.
As it can be quite a time consuming to reach a good overview of complex systems modeling in biomedical applications and research, the following review paper can help to breach the curing edge quickly:
https://www.researchgate.net/publication/330546521_Complex_Systems_and_Their_Use_in_Medicine_Concepts_Methods_and_Bio-Medical_Applications
With respect to ABM, its highest potential can be seen in the modelling of one-to-one interactions of cells, describing the behavior of the bacterial ecosystem living in our guts, which is closely related to the description of development and action of the immune system.
On a larger scale, ABM can be used to model the spread of infectious diseases, escape behavior, pedestrian and traffic flow, stock market, market, social interactions, sociology, politics, prediction of votes results, etc.
Good luck with your search and studies :-)
All who would like to know something about Agent-Based Modelling should put their hands on the following ABM programming environment.
From the author:
"NetLogo is a multi-agent programmable modeling environment. It is used by many tens of thousands of students, teachers and researchers worldwide. It also powers HubNet participatory simulations. It is authored by Uri Wilensky and developed at the CCL. You can download it free of charge. You can also try it online through NetLogo Web."
See more
https://ccl.northwestern.edu/netlogo/
There are many examples of various ABMs. It is really worth to go through all of them. On top of it, those examples can be simulated online using the web without even downloading and installing the software itself!
It is really worth to know and play with for a while. :-)
With an appropriately structured system, you can incorporate a wide range of model types in a system. My work has been on systems which are able to change the agents used to represent parts of a complex system in response to the conditions in the system and the requirements of the current set of agents.
Obvious candidates in industry might include using different assessment routines for collision detection and avoidance in autonomous vehicles based on their sensed location, the density of other vehicles and the nature of their cargo. A more intricate approach may include information passed between vehicles.
Most of my work has been in simulating the interactions between marine organisms/populations, human activity and industry, and management as a way of testing/comparing different management strategies ... I should think that there was a lot of scope for that kind of approach w.r.t. many of our responses to a changing and dynamic environment.
There are a number of "platforms" for ABM construction, but you really need to understand what sort of model you will be working on *before* you decide on implementation. I started my ABM work in 1992, at the time there weren't any readily identifiable platforms that were capable of doing the work, so I rolled my own. Having a common framework is a good thing (like standardised instruments), but there is a huge value to writing at least one ABM from scratch ... you develop a much deeper understanding of the connections, problems and benefits.
Of course your mileage may vary...
Cheers!
- Badges
- Science topic
- Similar topics
- Computer Science