Jonah McIntire of TNX Logistics and Anna Shaposhnikova of Transmetrics do an excellent job helping logistics and transportation teams understand what AI can do to help them. They outline two different AI approaches relevant to those industries and explain the difference between them so those evaluating AI software can make intelligent buying decisions.
Statistical AI, more often called machine learning (ML), relies on large volumes of historical and real-time data that contains information on, say, the processes associated with logistics or transportation. For instance, monitoring and managing a transportation system to find the most efficient routing for vehicles involves data that include route options available, the time required and fuel cost associated with various routes, and a host of other data elements. [1] ML algorithms predict the best choices, leaving the final decision-making to humans.
Using the same example (routing efficiency), AI planning algorithms are programmed to describe the “state of the world” as it presently exists. Giving those algorithms the rules that define allowable actions and the goals one wants to achieve enables the AI planning system to serve as an intelligent agent that offers advice.
As the authors point out, ML is best suited to problems of large and unstructured data, where greater experience is the primary way to improves outcomes. AI planning is better suited when the business has clear goals and a set of permissible decisions to achieve them.