This course is an introduction to mathematical models for automated decision-making. The domain of decision is basically a data environment that gives information (partly incomplete, partly redundant) on resources, actors, etc. that are subject to decisions aiming at some pre-defined goal(s).
Decision models can be just descriptive of the data environment, e.g., revealing hidden features of data such as trends, expected values, etc. Or they can be prescriptive, that is, encoide and then suggest decisions that adequately respect relations among resources and address specific goals. Or, finally, they can be predictive, in the sense of giving an outlook of the evolution of some domain data. Also, a decision model can incorporate multiple decision-makers that operate on the domain with possibly contrasting goals.
The course concentrates on those decision models that have the form of (integer) linear programming problems, describes the mathematical structure of those problems and gives indication of possible solution methods.
- Teacher: CLAUDIO ARBIB
- Non-editing teacher: STEFANO DI GENNARO