# PAMELI: A Meta-Algorithm for Computationally Expensive Multi-Objective Optimization Problems

**Authors:** Santiago Cuervo, Miguel Melgarejo, Angie Blanco-Cañon, Laura Reyes-Fajardo, Sergio Rojas-Galeano  
**arXiv:** [2103.10736](https://arxiv.org/abs/2103.10736)  
**Date:** 2021-03-19  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2103.10736v2/

## Key points

- The paper presents an algorithm for solving computationally expensive multi-objective optimization problems
- The algorithm uses surrogate models to evaluate only approximately Pareto-optimal solutions
- It performs a meta-search to identify optimal surrogate models and navigation strategies
- The adaptive search strategy allows the algorithm to adjust its approach based on new information
- Experimental comparison with a state-of-the-art surrogate-assisted evolutionary algorithm shows the effectiveness of the proposed PAMELI algorithm
- The algorithm efficiently finds approximate Pareto-optimal solutions
- It combines surrogate modeling with adaptive search strategies
- The results highlight its superiority over existing approaches
- It has potential in domains with limited or costly computational resources.

## Summary

The paper titled "PAMELI: A Meta-Algorithm for Computationally Expensive Multi-Objective Optimization Problems" presents an algorithm designed for solving multi-objective optimization problems that are computationally expensive. The proposed algorithm utilizes surrogate models to define a set of surrogate problems, allowing for the evaluation of only those solutions that are estimated to be approximately Pareto-optimal using the real expensive functions. In addition to searching for solutions, the algorithm also performs a meta-search to identify optimal surrogate models and navigation strategies for the optimization landscape. This adaptive search strategy enables the algorithm to adjust its approach based on new information obtained during the optimization process. To demonstrate the competitiveness of their approach, the authors conducted an experimental comparison with a state-of-the-art surrogate-assisted evolutionary algorithm on a set of benchmark problems. The results highlight the effectiveness of the proposed PAMELI algorithm in efficiently finding approximate Pareto-optimal solutions. Overall, this paper introduces a novel meta-algorithm that combines surrogate modeling with adaptive search strategies to tackle computationally expensive multi-objective optimization problems. The experimental results showcase its superiority over existing approaches and emphasize its potential in various domains where computational resources are limited or costly.