Advances in Optimization Algorithms

The field of optimization is experiencing significant advancements with the development of innovative algorithms and techniques. Recent research has focused on improving the efficiency and effectiveness of optimization methods, particularly in high-dimensional spaces and complex problem domains. A key trend is the incorporation of nature-inspired swarm intelligence and evolutionary algorithms, which have shown promising results in solving real-world optimization problems. Additionally, there is a growing interest in Bayesian optimization and trust region methods, which offer powerful tools for solving constrained optimization problems under limited evaluation budgets. These advancements have the potential to impact various fields, including engineering, energy optimization, and healthcare. Noteworthy papers in this area include:

  • The Feasibility-Driven Trust Region Bayesian Optimization algorithm, which rapidly refocuses its search to accelerate the discovery of feasible and good-quality solutions.
  • The Unbounded Differential Evolution framework, which eliminates replacement altogether and relies solely on selection mechanisms, enabling a more straightforward yet powerful search algorithm.
  • The Gene-invariant Optimal Mixing Evolutionary Algorithm, which preserves gene frequencies in the population and has shown striking potential in solving hierarchical problems.

Sources

Is the Fitness Dependent Optimizer Ready for the Future of Optimization?

Quasi-Monte Carlo hyperinterpolation

Is Selection All You Need in Differential Evolution?

Feasibility-Driven Trust Region Bayesian Optimization

The Pitfalls and Potentials of Adding Gene-invariance to Optimal Mixing

Estimate Hitting Time by Hitting Probability for Elitist Evolutionary Algorithms

Built with on top of