Logic-based decision support

This monograph is directly based on a series of ten lectures, of the same title, given at Rutgers University as part of the Advanced Research Institute on Discrete Applied Mathematics (ARIDAM). The lectures divide naturally into two parts. In Lectures 1 through 4, we cover the theory of representations for those problems which are solvable by mixed-integer programming (MIP), with emphasis on disjunctive formulation techniques. In Lectures 5 through 9, we discuss logic-based approaches to decision support which help to create more ‘intelligent’ systems. We try to show the huge potential for MIP techniques to assist in these approaches and, conversely, the potential for results in applied logic to be relevant in MIP research. Lecture 10 raises broader philosophical issues for speculation and discussion, and it attempts to put the work treated in the previous lectures in a broader perspective.