Decision Science – the next stage of Data Science for Project Management
Ajit Jaokar
Background Managing projects involves making decisions under uncertain conditions. Today, decision-making has become more involved with increasing business volatility. Current machine learning and deep learning techniques help to find insights from data. But businesses need to translate data insights into possible actionable decisions– often under uncertainty.
In other words, we need to change the emphasis by focusing on decisions as the unit of analysis. Decision science provides such a framework based on possible actions and the cost of the actions.
Hence, decision science could be the next stage of data science, acting as a missing link between insights and decisions – especially for complex project management.
Decision science:
Decision Science is the collection of quantitative techniques used to inform decision-making at the individual and population levels. It includes decision analysis, risk analysis, cost-benefit, and cost-effectiveness analysis, constrained optimization, simulation modeling, and behavioral decision theory, as well as parts of operations research, microeconomics, statistical inference, management control, cognitive and social psychology, and computer science.
Source: Harvard University - T H Chan school of Public Health
We could then ask: “How will current AI/ ML techniques evolve to encompass decision science strategies?”
Algorithms:
The ideas behind decision science are not new. Decision science is closely related to game theory and to other statistical and econometric methods. But, with the increasing availability of data, computation, and new algorithms, decision science is evolving.
Since decision science provides an evaluation of the optimal choices based on data, the results are more actionable, especially coupled with a degree of uncertainty associated with that action.
Decision science can be used in various domains for decision-making and project management, including manufacturing, health, environment, energy, etc. Many existing AI/ML algorithms could also be used in decision science, for example.
MCDA – Multi Criteria Decision Analysis
Reinforcement learning
Bayesian techniques for decisions under uncertainty
Natural Language Processing etc
Conclusion
At Digitty.io, we are working in this space to understand decision-making under conditions of uncertainty. I am especially interested in applying these ideas to complex problems where the risks are high. Traditional machine learning algorithms are concerned with minimization or maximization of a cost function. When you have project planning scenarios with multiple objectives, complex criteria, and dynamic scenarios, we need enhanced decision-making methods.
I welcome comments. We will share more about our work here.
About the author:
Ajit is the course director of the Oxford University course “Artificial Intelligence: Cloud and Edge Implementations”. Teaching this unique full-stack course gives Ajit an in-depth knowledge of the AI ecosystem.