Optimization Overview

Principles

Definition

Optimization is a methodological approach and a field of study focused on finding the best possible solution to a problem. It involves determining parameters that maximize or minimize a specified objective function while satisfying a set of constraints, aiming to enhance the performance, efficiency, or outcome of a system or process.

Ontological type

Core Methods

Mathematical Formulation

Solution Algorithms

Key developments

Key Figures

Conic-Interior Foundations era

J.F. Sturm[1] was active at the University of Groningen[2] and Erasmus University Rotterdam[3] during the Conic-Interior Foundations era (1987–2002). His key contributions include advancing optimization over symmetric cones and promoting practical tooling, notably SeDuMi 1.02[4], a Matlab toolbox described in the 1999 paper[4], which enabled scalable interior-point and conic approaches for large-scale problems. This era's methodological advances, driven by Sturm's work[1], supported rigorous relaxations, sensitivity analysis, and benchmarking that established reproducible performance baselines for conic solvers. These contributions by Sturm[1] were instrumental in enabling large-scale constrained nonlinear programming and in shaping the computational backbone for subsequent hybrid and discrete-continuous strategies.

Hybrid Decomposition & Matheuristics era

R. Venkata Rao [1] carried out work across Indian Institute of Technology Kanpur [3] and University of Maribor [4] during the Hybrid Decomposition & Matheuristics era. R. Venkata Rao [1] introduced Teaching–Learning-Based Optimization [7], a method for continuous non-linear large-scale problems and a foundational contribution to scalable hybrid solvers in this era. Lorenz T. Biegler [2] engaged with University of Pittsburgh [5] and Carnegie Mellon University [6] during this era. Lorenz T. Biegler [2] advanced large-scale nonlinear programming by integrating IPOPT into enterprise-wide dynamic optimization [8], a contribution connected with University of Pittsburgh [5] and Carnegie Mellon University [6] that underpinned scalable decomposition and matheuristic approaches.

Open-Source Solver Ecosystem era

Seyedali Mirjalili [1] emerges as a leading figure in open-source optimization during 2017–2023, with affiliations at The University of Queensland [2] and Yonsei University [3]. His paper The Arithmetic Optimization Algorithm [4] introduces a population-based metaheuristic inspired by arithmetic operations, offering a lightweight, adaptable tool that can be integrated into open-source modeling frameworks and reproducible workflows, thereby advancing cross-domain optimization in linear, nonlinear, and mixed-integer settings. This work [4], and Mirjalili's broader influence [1], catalyzed the reuse of metaheuristics within open-source solver ecosystems by providing a modular optimization primitive that supports rapid experimentation and method transfer. The University of Queensland [2] and Yonsei University [3] served as part of Mirjalili's base in this era, supporting a collaborative, open framework that democratized optimization across domains.