Comparer des méthodes

	Simheuristics	Optimisation stochastique
Domaine	Optimisation	Optimisation
Famille	Process / pipeline	Process / pipeline
Année d'origine	2015	1951 (SGD); 2014 (Adam)
Auteur d'origine	Juan et al.	—
Type	Hybrid simulation-optimization framework	Gradient-based iterative optimization
Source fondatrice	Juan, A. A., et al. (2015). A review of simheuristics: Extending metaheuristics to deal with stochastic combinatorial optimization problems. Operations Research Perspectives, 2, 62–72.	Robbins, H. & Monro, S. (1951). A Stochastic Approximation Method. Annals of Mathematical Statistics, 22(3), 400-407.
Alias	Simulation-based Metaheuristics, Stochastic Metaheuristics with Simulation, Hybrid Simulation-Optimization, Simülistik Sezgiseller	Stokastik Optimizasyon (SGD & Varyantları), stochastic gradient descent, SGD, Adam
Apparentées	3	3
Résumé	Simheuristics is a hybrid algorithmic framework that integrates Monte Carlo or discrete-event simulation into metaheuristic search procedures to solve stochastic combinatorial optimization problems. Introduced by Juan et al. in 2015, it addresses settings where objective function evaluations involve random variables, providing near-optimal solutions with probabilistic quality guarantees. The approach is especially suited for real-world logistics, transportation, and scheduling problems where uncertainty is inherent and classical deterministic solvers fail to capture variability.	Stochastic optimization is a family of iterative methods that minimize an objective function by computing gradients on randomly sampled subsets of data — mini-batches — rather than on the entire dataset at once. Pioneered by Robbins and Monro in 1951 as stochastic approximation, the approach became the standard engine for training large-scale machine-learning models through variants such as SGD with momentum, AdaGrad, RMSProp, and Adam.