internetboekhandel.nl
buttons Home pagina Kassa pagina, Winkelwagentje Contact info Email ons
leeg Home pagina Kassa pagina, Winkelwagentje Contact info Email ons Home pagina Rijks
Boekenweek
Rijks Home pagina Home pagina Kassa pagina, Winkelwagentje Contact info Email ons Besteller 60
 
Nederlands Buitenlands   Alles  Titel  Auteur  ISBN        
Technologie en beroepen
Energietechniek
Levertijd: 5 tot 11 werkdagen


Chang, Hyeong Soo

Simulation-based Algorithms for Markov Decision Processes

€ 168.25

Often, real-world problems modeled by Markov decision processes (MDPs) are difficult to solve in practise because of the curse of dimensionality. In others, explicit specifica


Taal / Language : English

Inhoudsopgave:
Selected Notation and Abbreviations xv
Markov Decision Processes
1(16)
Optimality Equations
3(2)
Policy Iteration and Value Iteration
5(2)
Rolling-horizon Control
7(1)
Survey of Previous Work on Computational Methods
8(3)
Simulation
11(3)
Preview of Coming Attractions
14(1)
Notes
15(2)
Multi-stage Adaptive Sampling Algorithms
17(44)
Upper Confidence Bound Sampling
19(21)
Regret Analysis in Multi-armed Bandits
19(1)
Algorithm Description
20(1)
Alternative Estimators
21(3)
Convergence Analysis
24(7)
Numerical Example
31(9)
Pursuit Learning Automata Sampling
40(19)
Algorithm Description
41(1)
Convergence Analysis
42(9)
Application to POMDPs
51(2)
Numerical Example
53(6)
Notes
59(2)
Population-based Evolutionary Approaches
61(28)
Evolutionary Policy Iteration
63(4)
Policy Switching
63(2)
Policy Mutation and Population Generation
65(1)
Stopping Rule
65(1)
Convergence Analysis
66(1)
Parallelization
67(1)
Evolutionary Random Policy Search
67(9)
Policy Improvement with Reward Swapping
68(3)
Exploration
71(2)
Convergence Analysis
73(3)
Numerical Examples
76(11)
A One-dimensional Queueing Example
76(9)
A Two-dimensional Queueing Example
85(2)
Extension to Simulation-based Setting
87(1)
Notes
87(2)
Model Reference Adaptive Search
89(60)
The Model Reference Adaptive Search Method
91(10)
The MRAS0 Algorithm (Idealized Version)
93(3)
The MRAS1 Algorithm (Adaptive Monte Carlo Version)
96(2)
The MRAS2 Algorithm (Stochastic Optimization)
98(3)
Convergence Analysis
101(28)
MRAS0 Convergence
101(6)
MRAS1 Convergence
107(9)
MRAS2 Convergence
116(13)
Application to MDPs via Direct Policy Learning
129(12)
Finite-horizon MDPs
130(1)
Infinite-horizon MDPs
130(2)
MDPs with Large State Spaces
132(1)
Numerical Examples
132(9)
Application to Infinite-horizon MDPs in Population-based Evolutionary Approaches
141(5)
Algorithm Description
141(2)
Numerical Examples
143(3)
Application to Finite-horizon MDPs Using Adaptive Sampling
146(2)
Notes
148(1)
On-line Control Methods via Simulation
149(28)
Simulated Annealing Multiplicative Weights Algorithm
153(12)
Basic Algorithm Description
154(1)
Convergence Analysis
155(3)
Convergence of the Sampling Version of the Algorithm
158(2)
Numerical Example
160(4)
Simulated Policy Switching
164(1)
Rollout
165(3)
Parallel Rollout
166(2)
Hindsight Optimization
168(6)
Numerical Example
169(5)
Notes
174(3)
Reference 177(10)
Index 187
Extra informatie: 
Hardback
189 pagina's
Januari 2007
439 gram
244 x 161 x 20 mm
SPRINGER NATURE gb

Levertijd: 5 tot 11 werkdagen



Andere titels binnen de rubriek:
Energietechniek