Deterministic global optimization with artificial neural networks embedded

DOI10.1007/s10957-018-1396-0zbMath1407.90263arXiv1801.07114OpenAlexW3103957401WikidataQ62736863 ScholiaQ62736863MaRDI QIDQ1730780

Artur M. Schweidtmann, Alexander Mitsos

Publication date: 6 March 2019

Published in: Journal of Optimization Theory and Applications (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1801.07114

zbMATH Keywords

multilayer perceptronmachine learningMcCormick relaxationssurrogate-based optimizationMAiNGO

Mathematics Subject Classification ID

Applications of mathematical programming (90C90) Nonconvex programming, global optimization (90C26) Nonlinear programming (90C30) General topics in artificial intelligence (68T01)

Related Items (18)

Obey validity limits of data-driven models through topological data analysis and one-class classification ⋮ Mixed-Integer Convex Nonlinear Optimization with Gradient-Boosted Trees Embedded ⋮ JANOS: An Integrated Predictive and Prescriptive Modeling Framework ⋮ Modeling design and control problems involving neural network surrogates ⋮ Surrogate-based branch-and-bound algorithms for simulation-based black-box optimization ⋮ Surrogate models of radiative transfer codes for atmospheric trace gas retrievals from satellite observations ⋮ Revisiting `survival of the fittest' principle in global stochastic optimisation: incorporating anisotropic mutations ⋮ Extended McCormick relaxation rules for handling empty arguments representing infeasibility ⋮ Comparison of MINLP formulations for global superstructure optimization ⋮ Convex and concave envelopes of artificial neural network activation functions for deterministic global optimization ⋮ Improved convex and concave relaxations of composite bilinear forms ⋮ Tighter McCormick relaxations through subgradient propagation ⋮ Strong mixed-integer programming formulations for trained neural networks ⋮ DeCODe: a community-based algorithm for generating high-quality decompositions of optimization problems ⋮ Deterministic global optimization with Gaussian processes embedded ⋮ Data-driven spatial branch-and-bound algorithms for box-constrained simulation-based optimization ⋮ Working fluid selection for organic rankine cycles via deterministic global optimization of design and operation ⋮ A two-stage exact algorithm for optimization of neural network ensemble

Uses Software

• SCIP
• CONOPT
• CPLEX
• FADBAD++
• DICOPT
• filib++
• libMC
• ANTIGONE
• NLopt
• MC++
• MAiNGO
• DFVLR-SQP
• PRMLT

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Reverse propagation of McCormick relaxations
• Differentiable McCormick relaxations
• Three enhancements for optimization-based bound tightening
• Convergence analysis of multivariate McCormick relaxations
• Convergence rate of McCormick relaxations
• Generalized McCormick relaxations
• A reduced space branch and bound algorithm for global optimization.
• Deterministic global optimization of process flowsheets in a reduced space using McCormick relaxations
• Corrections to: ``Differentiable McCormick relaxations
• Multilayer feedforward networks are universal approximators
• A polyhedral branch-and-cut approach to global optimization
• Tighter McCormick relaxations through subgradient propagation
• ANTIGONE: algorithms for coNTinuous/Integer global optimization of nonlinear equations
• Multivariate McCormick relaxations
• Erratum to: ``Multivariate McCormick relaxations
• Convex and concave relaxations of implicit functions
• FILIB++, a fast interval library supporting containment computations
• McCormick-Based Relaxations of Algorithms
• Computability of global solutions to factorable nonconvex programs: Part I — Convex underestimating problems
• CONOPT—A Large-Scale GRG Code

This page was built for publication: Deterministic global optimization with artificial neural networks embedded