Guided Randomness in Optimization, Volume 1 [Hardcover]

The performance of an algorithm used depends on the GNA. This book focuses on the comparison of optimizers, it defines a stress-outcome approach which can be derived all the classic criteria (median, average, etc.) and other more sophisticated.   Source-codes used for the examples are also presented, this allows a reflection on the superfluous chance, succinctly explaining why and how the stochastic aspect of optimization could be avoided in some cases.

INTRODUCTION xv

PART 1. RANDOMNESS IN OPTIMIZATION  1

CHAPTER 1. NECESSARY RISK 3

1.1. No better than random search 3

1.1.1. Uniform random search 4

1.1.2. Sequential search 5

1.1.3. Partial gradient 5

1.2. Better or worse than random search 7

1.2.1. Positive correlation problems 8

1.2.2. Negative correlation problems 10

CHAPTER 2. RANDOM NUMBER GENERATORS (RNGS) 13

2.1. Generator types 14

2.2. True randomness 15

2.3. Simulated randomness 15

2.3.1. KISS 16

2.3.2. Mersenne-Twister 16

2.4. Simplified randomness 17

2.4.1. Linear congruential generators 18

2.4.2. Additive 20

2.4.3. Multiplicative 22

2.5. Guided randomness 24

2.5.1. Gaussian 24

2.5.2. Bell 24

2.5.3. Cauchy 27

2.5.4. Lévy 28

2.5.5. Log-normal 28

2.5.6. Composite distributions 28

CHAPTER 3. THE EFFECTS OF RANDOMNESS 33