Glossary#

Key terms and concepts used in the Evolver project.

Base-level Metaheuristic#: Algorithm (e.g., NSGA-II, MOEA/D) being configured through meta-optimization.
Meta-optimization#: Process of finding optimal configurations for base-level metaheuristics.
Solution Encoding#: Representation of candidate solutions (e.g., binary, real, permutation).
Parameter Space#: All possible configurations for a metaheuristic, including their constraints.
Quality Indicator#: Metric for evaluating metaheuristic performance (e.g., Hypervolume, Epsilon).
Reference Front#: Best-known approximation to the true Pareto front for a problem.
Non-dominated Solution#: Solution not worse than any other in all objectives and better in at least one.
Pareto Front#: Set of all non-dominated solutions in the objective space.
Hypervolume (HV)#: Volume of objective space dominated by a solution set.
Epsilon Indicator#: Distance needed to make one solution set dominate another.
IGD#: Average distance from reference front to nearest solution.
MOEA#: Multi-Objective Evolutionary Algorithm.
NSGA-II#: Non-dominated Sorting Genetic Algorithm II.
SMPSO#: Speed-constrained Multi-objective PSO.
RDEMOEA#: Ranking and Density Estimator Multi-Objective Evolutionary Algorithm.
TSP#: Traveling Salesman Problem.
ZDT/DTLZ#: Benchmark problem sets for multi-objective optimization.
JVM#: Java Virtual Machine.
YAML#: Human-readable data format for configuration.
API#: Application Programming Interface.
Javadoc#: Documentation generator for Java code.
Sphinx#: Documentation generator for Python/other languages.
RST#: reStructuredText markup language.
CI/CD#: Continuous Integration/Continuous Deployment.
Unit Test#: Testing individual components in isolation.
Integration Test#: Testing combined components as a group.
Fitness Function#: Evaluates solution quality in optimization.
Crossover#: Genetic operator combining parent solutions.
Mutation#: Random modification of solutions.
Selection#: Choosing solutions for next generation.
Population#: Set of candidate solutions.
Generation#: Single iteration in evolutionary algorithms.
Convergence#: Algorithm approaching a stable solution.
Diversity#: Variation among solutions in population.
Local Search#: Improving solutions through small changes.
Global Search#: Exploring the entire solution space.
Heuristic#: Problem-solving approach for approximate solutions.
Metaheuristic#: High-level strategy for heuristic optimization.
Parallelization#: Executing multiple computations simultaneously.
Scalability#: Performance with increasing problem size.
Robustness#: Consistent performance across different scenarios.
Parameter Tuning#: Optimizing algorithm parameters.
Multi-objective Optimization#: Optimizing multiple conflicting objectives.
Single-objective Optimization#: Optimizing a single objective.
Constraint Handling#: Managing solution feasibility.
Feasible Solution#: Solution satisfying all constraints.
Pareto Optimality#: A solution is Pareto optimal if there exists no other feasible solution that would improve one objective without simultaneously worsening at least one other objective. The Pareto optimal solutions form the Pareto front in the objective space.
Dominance#: One solution being better than another in all objectives.
Non-dominated Sorting#: Ranking solutions by dominance levels.
Crowding Distance#: Density estimation in objective space.
Archive#: Storage for non-dominated solutions.
PSO#: Particle Swarm Optimization.
DE#: Differential Evolution.