MHMetLib: A New Performance Indicators Library for Trajectory and Population-Based Metaheuristics

2026 • Juan Pablo Cassis, Joaquín Gatica, Nicolás Rojas-Morales, Carlos Hernández Castellanos

Traditional performance evaluation metrics for metaheuristics focus primarily on the best solution obtained or the computational resources required to solve a problem. However, further valuable insights into algorithmic components and design can be extracted during the search process. This information enhances understanding of algorithmic design, supports algorithm configuration, and informs methodological advancements. This study presents MHMetLib, a unified indicators library for analyzing both trajectory-based and population-based metaheuristics. The library offers indicators for convergence, diversity, and operator behavior. This enables comprehensive behavioral analysis beyond traditional scalar measures. Each indicator is formally defined and details its objective, computational requirements, and application domain. MHMetLib was implemented and evaluated within the IOHProfiler platform. This …

Understanding ant colony optimization search through local optima networks

2026 • Nicolás Rojas-Morales, Elizabeth Montero, Leslie Pérez, Gabriela Ochoa, María Cristina Riff

AbstractThe behavior of Ant Colony Optimization (ACO) algorithms is based on a collective learning process defined by a pheromone mechanism. This process is difficult to understand due to the scheduling of pheromone deposition and evaporation, the influence of parameter values, the bias in solution construction, and the problem instance size. This work aims to extend Local Optima Networks (LONs) to understand the path traversed by the pheromone learning mechanism in the fitness landscape described by ACO algorithms. Our ant-based LONs incorporate a definition of network edges that expresses the persistence of the pheromone influence in the search process. Also, we study a simplified network, Deposit LON, that contains only nodes that deposit pheromones. We aim to analyze a population-based search algorithm’s exploitation and exploration behavior through its network features. We evaluate our …

2SLG: A Metaheuristic Approach for Procedural Level Generation for 2-D Platformers

2025 • Carlos Saavedra Díaz, Nicolas Rojas-Morales

Procedural generation of 2-D platform game levels aims to automate the design process, enabling dynamic content creation. However, generating levels that are not only feasible but also engaging and aesthetically coherent remains a complex and open challenge. We propose 2SLG: a 2-step metaheuristic approach that procedurally generates feasible and playable 2-D platform game levels. The first step consists of an Evolutionary Algorithm that places the platforms on the level, followed by a Hill Climbing step, for placing the enemies and obstacles; both steps aim to generate levels that satisfy user-defined requirements while maintaining aesthetic quality. To evaluate the performance of the algorithm, experiments were performed, generating Super Mario Bros levels, changing features such as the expected value of platforms, enemies, and linearity. The results show that the proposal can effectively generate …

CLIRACE

2025 • Diana Gil-Silvestre, Nicolás Rojas-Morales

H-IRACE

2025 • Diana Gil-Silvestre, Nicolás Rojas-Morales

MHMetLib

2025 • Juan Pablo Cassis, Joaquin Gatica, Nicolás Rojas-Morales, Carlos Hernandez