Heuristic Techniques for Stochastic Combinatorial Optimization
Sprache des Titels:
The tutorial starts with a very short recapitulation of the most important types of stochastic combinatorial optimization (SCO) problems, as they occur in production, logistics, scheduling, facility location, energy, environment, healthcare management, telecommunication and other areas. Since in many cases, already the deterministic counterparts of these problems are NP-hard and can often not be solved exactly, it is no surprise that also in the SCO domain, problem instances of realistic size and complexity frequently require heuristic solution techniques. The tutorial addresses some approximate or heuristic SCO solution approaches either derived from exact methods (as branch-and-bound or progressive hedging) or from prominent metaheuristics (as variants of local search, simulated annealing, evolutionary algorithms, or swarm intelligence algorithms). Special emphasis is given to a discussion of fixed-sample and variable-sample Monte Carlo techniques in SCO. Known results on analytical properties of the considered SCO algorithms are indicated. Finally, a short outline of multi-objective SCO problems and their solution is provided.
Ändere diese SeiteQuelleEntfernenUmbenennenBerechtigungenHistorie Kommentieren Datei anhängen