AI-ordlista
Den kompletta ordlistan över AI
Evolutionary Algorithm
Metaheuristic inspired by biological evolution using genetic operators such as selection, crossover, and mutation to explore the search space of optimal solutions.
Hypervolume
Performance indicator measuring the volume of the objective space dominated by a set of non-dominated solutions, evaluating both convergence and diversity on the Pareto front.
Multi-Objective Policy Gradient
Reinforcement learning method that directly optimizes policy parameters to maximize a multi-objective reward vector using stochastic gradient techniques.
Evolutionary Archive
Data structure that preserves non-dominated solutions discovered during evolution, used to maintain the history of Pareto front approximations and guide future search.
Multi-Objective Adaptive Mutation
Mutation operator whose parameters adapt dynamically based on population diversity and convergence to balance exploration and exploitation in the multi-objective search space.
Multi-Objective Tournament Selection
Selection mechanism in evolutionary algorithms where candidates are compared based on dominance or diversity criteria to determine individuals participating in reproduction.
Scalarizing Aggregation Function
Technique that transforms a multi-objective problem into a single-objective problem by linearly or non-linearly combining different objectives with variable weighting coefficients.
Evolutionary Diversity
Ability of an evolutionary algorithm to maintain a balanced distribution of solutions on the Pareto front, avoiding premature convergence to localized regions of the search space.
Evolutionary Convergence
The tendency of an evolutionary algorithm to approach the true Pareto front over generations, measured by the distance between obtained solutions and optimal non-dominated solutions.