Abstract:In real-life decision analysis, significant recognition has been given to theunrealistic expectation of numerically precise information. Many modernapproaches attempting to handle imprecision have focused more on representationand less on evaluation. The DELTA method, as one of the fewleading approaches, challenges this issue by its evaluation framework thatcan accommodate both precision and imprecision. However, computationally,DELTA and similar approaches may incur time-consuming calculationsdue to the introduction of imprecise information concerning probability andutility. In general, those problems are bounded non-convex bilinear optimizationprograms with disjoint linear constraints, which cannot be solvedeffectively by any existing general-purpose global optimization technique.This thesis presents two enhanced cutting plane algorithms for solvingbounded disjoint bilinear programs arising in computational decision analysis.Each algorithm consists of a local phase designed to determine a localoptimizer from an approximate solution, and a global phase designed to systematicallyexplore the feasible region, subset by subset. These two phasesare switched automatically during the global search procedure. The basicframework builds upon previously developed efficient cutting plane methods.By embedding the lower bounding technique in a branch and bound procedure,the improvement of their performances seems encouraging in the lightof computational experience. Even though the motivation to develop thesealgorithms stems from computational decision analysis, the idea can also beextended to the development of optimization approaches for handling generalbounded disjoint bilinear programs, especially for larger sized ones.

Continuous cutting plane algorithms in integer programming

A Difference-of-Convex Cutting Plane Algorithm for Mixed-Binary Linear Program

Learning to Remove Cuts in Integer Linear Programming

Differentiable Cutting-plane Layers for Mixed-integer Linear Optimization

A Disjunctive Cutting Plane Algorithm for Bilinear Programming

Machine Learning for Cutting Planes in Integer Programming: A Survey

Continuous Piecewise Linear Programming Via Concave Optimization and Genetic Algorithm

Accelerating Convergence of Cutting Plane Algorithms for Disjoint Bilinear Programming

Learning Cut Generating Functions for Integer Programming

Machine Learning for Optimization-Based Separation of Mixed-Integer Rounding Cuts

A New Algorithm for State-Constrained Separated Continuous Linear Programs

Cutting planes for signomial programming

A Context-Aware Cutting Plane Selection Algorithm for Mixed-Integer Programming

A Finitely Convergent Cutting Plane, and a Bender's Decomposition Algorithm for Mixed-Integer Convex and Two-Stage Convex Programs using Cutting Planes

A computational study of cutting-plane methods for multi-stage stochastic integer programs

A cutting-plane method to nonsmooth multiobjective optimization problems

Improved Cutting Plane Methods for Disjoint Bilinear Programming

LLMs for Cold-Start Cutting Plane Separator Configuration

Cut-based Conflict Analysis in Mixed Integer Programming

Learning to Configure Separators in Branch-and-Cut

Structural Analysis of Branch-and-Cut and the Learnability of Gomory Mixed Integer Cuts