HOME > 상세정보

상세정보

Optimal learning

Optimal learning

자료유형
단행본
개인저자
Powell, Warren B., 1955-. Ryzhov, Ilya Olegovich, 1985-.
서명 / 저자사항
Optimal learning / Warren B. Powell, Operations Research and Financial Engineering, Princeton University, Ilya O. Ryzhov, Robert H. Smith School of Business, University of Maryland.
발행사항
Hoboken, New Jersey : Wiley, c2012.
형태사항
xix, 384 p. : ill. ; 25 cm.
총서사항
Wiley series in probability and statistics
ISBN
9780470596692 (hardback)
요약
"This text presents optimal learning techniques with applications in energy, homeland security, health, sports, transportation science, biomedical research, biosurveillance, stochastic optimization, high technology, and complex resource allocation problems. The coverage utilizes a relatively new class of algorithmic strategies known as approximate dynamic programming, which merges dynamic programming (Markov decision processes), math programming (linear, nonlinear, and integer), simulation, and statistics. It features mathematical techniques that are applicable to a variety of situations, from identifying promising drug candidates to figuring out the best evacuation plan in the event of a natural disaster"--
서지주기
Includes bibliographical references and index.
일반주제명
Machine learning.
000 00000cam u2200205 a 4500
001 000045708617
005 20190226150433
008 120611s2012 njua b 001 0 eng
010 ▼a 2011047629
020 ▼a 9780470596692 (hardback)
035 ▼a (KERIS)REF000016618829
040 ▼a DLC ▼b eng ▼c DLC ▼e rda ▼d 211009
050 0 0 ▼a Q325.5 ▼b .P69 2012
082 0 0 ▼a 006.3/1 ▼2 23
084 ▼a 006.31 ▼2 DDCK
090 ▼a 006.31 ▼b P886o
100 1 ▼a Powell, Warren B., ▼d 1955-.
245 1 0 ▼a Optimal learning / ▼c Warren B. Powell, Operations Research and Financial Engineering, Princeton University, Ilya O. Ryzhov, Robert H. Smith School of Business, University of Maryland.
260 ▼a Hoboken, New Jersey : ▼b Wiley, ▼c c2012.
300 ▼a xix, 384 p. : ▼b ill. ; ▼c 25 cm.
490 1 ▼a Wiley series in probability and statistics
504 ▼a Includes bibliographical references and index.
520 ▼a "This text presents optimal learning techniques with applications in energy, homeland security, health, sports, transportation science, biomedical research, biosurveillance, stochastic optimization, high technology, and complex resource allocation problems. The coverage utilizes a relatively new class of algorithmic strategies known as approximate dynamic programming, which merges dynamic programming (Markov decision processes), math programming (linear, nonlinear, and integer), simulation, and statistics. It features mathematical techniques that are applicable to a variety of situations, from identifying promising drug candidates to figuring out the best evacuation plan in the event of a natural disaster"-- ▼c Provided by publisher.
650 0 ▼a Machine learning.
700 1 ▼a Ryzhov, Ilya Olegovich, ▼d 1985-.
830 0 ▼a Wiley series in probability and statistics.
945 ▼a KLPA

No. 소장처 청구기호 등록번호 도서상태 반납예정일 예약 서비스
No. 1 소장처 중앙도서관/서고6층/ 청구기호 006.31 P886o 등록번호 111668060 도서상태 대출가능 반납예정일 예약 서비스 B M
No. 2 소장처 과학도서관/Sci-Info(2층서고)/ 청구기호 006.31 P886o 등록번호 121248054 도서상태 대출가능 반납예정일 예약 서비스 B M
No. 소장처 청구기호 등록번호 도서상태 반납예정일 예약 서비스
No. 1 소장처 중앙도서관/서고6층/ 청구기호 006.31 P886o 등록번호 111668060 도서상태 대출가능 반납예정일 예약 서비스 B M
No. 소장처 청구기호 등록번호 도서상태 반납예정일 예약 서비스
No. 1 소장처 과학도서관/Sci-Info(2층서고)/ 청구기호 006.31 P886o 등록번호 121248054 도서상태 대출가능 반납예정일 예약 서비스 B M

컨텐츠정보

저자소개

Warren B. Powell(지은이)

Ilya O. Ryzhov(지은이)

정보제공 : Aladin

목차

Preface	p. xv
Acknowledgments	p. xix
1	The Challenges of Learning	p. 1
1.1	    Learning the Best Path	p. 2
1.2	    Areas of Application	p. 4
1.3	    Major Problem Classes	p. 12
1.4	    The Different Types of Learning	p. 13
1.5	    Learning from Different Communities	p. 16
1.6	    Information Collection Using Decision Trees	p. 18
1.6.1	        A Basic Decision Tree	p. 18
1.6.2	        Decision Tree for Offline Learning	p. 20
1.6.3	        Decision Tree for Online Learning	p. 21
1.6.4	        Discussion	p. 25
1.7	    Website and Downloadable Software	p. 26
1.8	    Goals of this Book	p. 26
            Problems	p. 27
2	Adaptive Learning	p. 31
2.1	    The Frequentist View	p. 32
2.2	    The Bayesian View	p. 33
2.2.1	        The Updating Equations for Independent Beliefs	p. 34
2.2.2	        The Expected Value of Information	p. 36
2.2.3	        Updating for Correlated Normal Priors	p. 38
2.2.4	        Bayesian Updating with an Uninformative Prior	p. 41
2.3	    Updating for Non-Gaussian Priors	p. 42
2.3.1	        The Gamma-Exponential Model	p. 43
2.3.2	        The Gamma-Poisson Model	p. 44
2.3.3	        The Pareto-Uniform Model	p. 45
2.3.4	        Models for Learning Probabilities*	p. 46
2.3.5	        Learning an Unknown Variance*	p. 49
2.4	    Monte Carlo Simulation	p. 51
2.5	    Why Does It Work?*	p. 54
2.5.1	        Derivation of ¿	p. 54
2.5.2	        Derivation of Bayesian Updating Equations for Independent Beliefs	p. 55
2.6	    Bibliographic Notes	p. 57
            Problems	p. 57
3	The Economics of Information	p. 61
3.1	    An Elementary Information Problem	p. 61
3.2	    The Marginal Value of Information	p. 65
3.3	    An information Acquisition Problem	p. 68
3.4	    Bibliographic Notes	p. 70
            Problems	p. 70
4	Ranking and Selection	p. 71
4.1	    The Model	p. 72
4.2	    Measurement Policies	p. 75
4.2.1	        Deterministic Versus Sequential Policies	p. 75
4.2.2	        Optimal Sequential Policies	p. 76
4.2.3	        Heuristic Policies	p. 77
4.3	    Evaluating Policies	p. 81
4.4	    More Advanced Topics*	p. 83
4.4.1	        An Alternative Representation of the Probability Space	p. 83
4.4.2	        Equivalence of Using True Means and Sample Estimates	p. 84
4.5	    Bibliographic Notes	p. 85
            Problems	p. 85
5	The Knowledge Gradient	p. 89
5.1	    The Knowledge Gradient for Independent Beliefs	p. 90
5.1.1	        Computation	p. 91
5.1.2	        Some Properties of the Knowledge Gradient	p. 93
5.1.3	        The Four Distributions of Learning	p. 94
5.2	    The Value of Information and the S-Curve Effect	p. 95
5.3	    Knowledge Gradient for Correlated Beliefs	p. 98
5.4	    Anticipatory Versus Experiential Learning	p. 103
5.5	    The Knowledge Gradient for Some Non-Gaussian Distributions	p. 105
5.5.1	        The Gamma-Exponential Model	p. 105
5.5.2	        The Gamma-Poisson Model	p. 108
5.5.3	        The Pareto-Uniform Model	p. 109
5.5.4	        The Beta-Bernoulli Model	p. 111
5.5.5	        Discussion	p. 113
5.6	    Relatives of the Knowledge Gradient	p. 114
5.6.1	        Expected Improvement	p. 114
5.6.2	        Linear Loss*	p. 115
5.7	    The Problem of Priors	p. 118
5.8	    Discussion	p. 120
5.9	    Why Does It Work?*	p. 120
5.9.1	        Derivation of the Knowledge Gradient Formula	p. 120
5.10	    Bibliographic Notes	p. 125
            Problems	p. 125
6	Bandit Problems	p. 139
6.1	    The Theory and Practice of Gittins Indices	p. 141
6.1.1	        Gittins Indices in the Beta-Bernoulli Model	p. 142
6.1.2	        Gittins Indices in tie Normal-Normal Model	p. 145
6.1.3	        Approximating Gittins Indices	p. 147
6.2	    Variations of Bandit Problems	p. 148
6.3	    Upper Confidence Bounding	p. 149
6.4	    The Knowledge Gradient for Bandit Problems	p. 151
6.4.1	        The Basic Idea	p. 151
6.4.2	        Some Experimental Comparisons	p. 153
6.4.3	        Non-Normal Models	p. 156
6.5	    Bibliographic Notes	p. 157
            Problems	p. 157
7	Elements of a Learning Problem	p. 163
7.1	    The States of our System	p. 164
7.2	    Types of Decisions	p. 166
7.3	    Exogenous Information	p. 167
7.4	    Transition Functions	p. 168
7.5	    Objective Functions	p. 168
7.5.1	        Designing Versus Controlling	p. 169
7.5.2	        Measurement Costs	p. 170
7.5.3	        Objectives	p. 170
7.6	    Evaluating Policies	p. 175
7.7	    Discussion	p. 177
7.8	    Bibliographic Notes	p. 178
            Problems	p. 178
8	Linear Belief Models	p. 181
8.1	    Applications	p. 182
8.1.1	        Maximizing Ad Clicks	p. 182
8.1.2	        Dynamic Pricing	p. 184
8.1.3	        Housing Loans	p. 184
8.1.4	        Optimizing Dose Response	p. 185
8.2	    A Brief Review of Linear Regression	p. 186
8.2.1	        The Normal Equations	p. 186
8.2.2	        Recursive Least Squares	p. 187
8.2.3	        A Bayesian Interpretation	p. 188
8.2.4	        Generating a Prior	p. 189
8.3	    The Knowledge Gradient for a Linear Model	p. 191
8.4	    Application to Drug Discovery	p. 192
8.5	    Application to Dynamic Pricing	p. 196
8.6	    Bibliographic Notes	p. 200
            Problems	p. 200
9	Subset Selection Problems	p. 203
9.1	    Applications	p. 205
9.2	    Choosing a Subset Using Ranking and Selection	p. 207
9.2.1	        Setting Prior Means and Variances	p. 207
9.2.2	        Two Strategies for Setting Prior Covariances	p. 208
9.3	    Larger Sets	p. 209
9.3.1	        Using Simulation to Reduce the Problem Size	p. 210
9.3.2	        Computational Issues	p. 212
9.3.3	        Experiments	p. 213
9.4	    Very Large Sets	p. 214
9.5	    Bibliographic Notes	p. 216
            Problems	p. 216
10	Optimizing a Scalar Function	p. 219
10.1	    Deterministic Measurements	p. 219
10.2	    Stochastic Measurements	p. 223
10.2.1	        The Model	p. 223
10.2.2	        Finding the Posterior Distribution	p. 224
10.2.3	        Choosing the Measurement	p. 226
10.2.4	        Discussion	p. 229
10.3	    Bibliographic Notes	p. 229
            Problems	p. 229
11	Optimal Bidding	p. 231
11.1	    Modeling Customer Demand	p. 233
11.1.1	        Some Valuation Models	p. 233
11.1.2	        The Logit Model	p. 234
11.2	    Bayesian Modeling for Dynamic Pricing	p. 237
11.2.1	        A Conjugate Prior for Choosing Between Two Demand Curves	p. 237
11.2.2	        Moment Matching for Nonconjugate Problems	p. 239
11.2.3	        An Approximation for the Logit Model	p. 242
11.3	    Bidding Strategies	p. 244
11.3.1	        An Idea From Multi-Armed Bandits	p. 245
11.3.2	        Bayes-Greedy Bidding	p. 245
11.3.3	        Numerical Illustrations	p. 247
11.4	    Why Does It Work?*	p. 251
11.4.1	        Moment Matching for Pareto Prior	p. 251
11.4.2	        Approximating the Logistic Expectation	p. 252
11.5	    Bibliographic Notes	p. 253
            Problems	p. 254
12	Stopping Problems	p. 255
12.1	    Sequential Probability Ratio Test	p. 255
12.2	    The Secretary Problem	p. 261
12.2.1	        Setup	p. 261
12.2.2	        Solution	p. 262
12.3	    Bibliographic Notes	p. 266
            Problems	p. 266
13	Active Learning in Statistics	p. 269
13.1	    Deterministic Policies	p. 270
13.2	    Sequential Policies for Classification	p. 274
13.2.1	        Uncertainty Sampling	p. 274
13.2.2	        Query by Committee	p. 275
13.2.3	        Expected Error Reduction	p. 277
13.3	    A Variance-Minimizing Policy	p. 277
13.4	    Mixtures of Gaussians	p. 280
13.4.1	        Estimating Parameters	p. 280
13.4.2	        Active Learning	p. 282
13.5	    Bibliographic Notes	p. 283
14	Simulation Optimization	p. 285
14.1	    Indifference Zone Selection	p. 288
14.1.1	        Batch Procedures	p. 288
14.1.2	        Sequential Procedures	p. 290
14.1.3	        The 0-1 Procedure: Connection to Linear Loss	p. 292
14.2	    Optimal Computing Budget Allocation	p. 293
14.2.1	        Indifference-Zone Version	p. 293
14.2.2	        Linear Loss Version	p. 295
14.2.3	        When Does It Work?	p. 295
14.3	    Model-Based Simulated Annealing	p. 296
14.4	    Other Areas of Simulation Optimization	p. 298
14.5	    Bibliographic Notes	p. 299
15	Learning in Mathematical Programming	p. 301
15.1	    Applications	p. 303
15.1.1	        Piloting a Hot Air Balloon	p. 303
15.1.2	        Optimizing a Portfolio	p. 308
15.1.3	        Network Problems	p. 309
15.1.4	        Discussion	p. 313
15.2	    Learning on Graphs	p. 313
15.3	    Alternative Edge Selection Policies	p. 317
15.4	    Learning Costs for Linear Programs*	p. 318
15.5	    Bibliographic Notes	p. 324
16	Optimizing Over Continuous Measurements	p. 325
16.1	    The Belief Model	p. 327
16.1.1	        Updating Equations	p. 328
16.1.2	        Parameter Estimation	p. 330
16.2	    Sequential Kriging Optimization	p. 332
16.3	    The Knowledge Gradient for Continuous Parameters*	p. 334
16.3.1	        Maximizing the Knowledge Gradient	p. 334
16.3.2	        Approximating the Knowledge Gradient	p. 335
16.3.3	        The Gradient of the Knowledge Gradient	p. 336
16.3.4	        Maximizing the Knowledge Gradient	p. 338
16.3.5	        The KGCP Policy	p. 339
16.4	    Efficient Global Optimization	p. 340
16.5	    Experiments	p. 341
16.6	    Extension to Higher-Dimensional Problems	p. 342
16.7	    Bibliographic Notes	p. 343
17	Learning With a Physical State	p. 345
17.1	    Introduction to Dynamic Programming	p. 347
17.1.1	        Approximate Dynamic Programming	p. 348
17.1.2	        The Exploration vs. Exploitation Problem	p. 350
17.1.3	        Discussion	p. 351
17.2	    Some Heuristic Learning Policies	p. 352
17.3	    The Local Bandit Approximation	p. 353
17.4	    The Knowledge Gradient in Dynamic Programming	p. 355
17.4.1	        Generalized Learning Using Basis Functions	p. 355
17.4.2	        The Knowledge Gradient	p. 358
17.4.3	        Experiments	p. 361
17.5	    An Expected Improvement Policy	p. 363
17.6	    Bibliographic Notes	p. 364
Index	p. 381

관련분야 신착자료

Taulli, Tom (2020)