Maybe it could be interesting to gather as many as possible (measureable, countable) information about a big number of projects (e.g. team size, company size, budget, number of documents, country, number of team meetings, etc, etc) and train a neuronal network with this data.
Output could be expected duration of the project and fail/success
I do not know the subject and title of your research, But I suggest that in your research, you use an integrated method based on fuzzy network analysis and multi-fuzzy multi-criteria decision making to examine one of the desired management issues.
You might be surprised, but what I published is related to my experience in project management where AI played a major role during teleworking experimentation in 2005 where the participants were disseminated in 12 countries, pls read this article carefully
Yes I have article for SEL program inteligence thinking were 2 title "
1) Sutinan Pukdeewut and Chalard Chantarasombat. Creative Thinking Development Program for Learning Activity Management of Secondary School Teachers,, Proceedings of IFKAD ISSN 2280-787X, 8 th International Forum on Knowledge Management Asset Dynamics : Smart Growth : Organizations, Cities and Communities, Institute of Knowledge Asset Management, Zagreb, Croatia, pg, 1112-1124, 12-14 June 2013,
2) Sutinan Pukdeewut, Chalard Chantarasombat & Pattananusorn Satapornwong “Creative Thinking Development Program for Learning Activity Management of Secondary School Teachers,” International Education Studies. Vol.6. No.12 ; 2013. Online Published : 28 November, 2013.
You might find something of interest in my google scholar page: https://scholar.google.com/citations?user=XOg2QnoAAAAJ&hl=en
In the late 80s and early 90s "scheduling" was considered a field of artificial intelligence. This system that I built for NASA to shcedule the operations of Space Shuttle missions, at the time, was considered artificial intelligence,. And temporal reasoning (in my dissertation) would be considered "domain specific artificial intelligence" Genetic algorithms used to improve the quality of project schedules would be considered artificial intelligence, and benchmarking alternative AI algorithms for scheduling would be considered a contribution to the field of artificial intelligence. Let me know if you would like to discuss and I will schedule a google-meeting.