CHAPTER 12 STATISTICAL METHODS FOR OPTIMIZATION IN DISCRETE PROBLEMS

1
CHAPTER 12 STATISTICAL METHODS FOR OPTIMIZATION
IN DISCRETE PROBLEMS
Slides for Introduction to Stochastic Search and
Optimization (ISSO) by J. C. Spall

• Organization of chapter in ISSO
• Basic problem in multiple comparisons
• Finite number of elements in search domain ?
• Tukey-Kramer test
• Many-to-one tests for sharper analysis
• Measurement noise variance known
• Measurement noise variance unknown (estimated)
• Ranking and selection methods

2
Background

• Statistical methods used here to solve
  optimization problem
• Not just for evaluation purposes
• Extending standard pairwise t-test to multiple
  comparisons
• Let ? ? ? ? ?1, ?2, , ?K be finite search
  space (K possible options)
• Optimization problem is to find the j such that
  ?? ?j
• Only have noisy measurements of L(?i)

3
Applications with Monte Carlo Simulations

• Suppose wish to evaluate K possible options in a
  real system
• Too difficult to use real system to evaluate
  options
• Suppose run Monte Carlo simulation(s) for each of
  the K options
• Compare options based on a performance measure
  (or loss function) L(?) representing average
  (mean) performance
• ? represents options that can be varied
• Monte Carlo simulations produce noisy measurement
  of loss function L at each option

4
Statistical Hypothesis Testing

• Null hypothesis All options in ? ? ?1, ?2, ,
  ?K are effectively the same in the sense that
  L(?1) L(?2) L(?K)
• Challenge in multiple comparisons alternative
  hypothesis is not unique
• Contrasts with standard pairwise t-test
• Analogous to standard t-test, hypothesis testing
  based on collecting sample values of L(?1),
  L(?2), and L(?K), forming sample means

5
TukeyKramer Test

• Tukey (1953) and Kramer (1956) independently
  developed popular multiple comparisons analogue
  to standard t-test
• Recall null hypothesis that all options in ? ?
  ?1, ?2, , ?K are effectively the same in the
  sense that L(?1) L(?2) L(?K)
• TukeyKramer test forms multiple acceptance
  intervals for K(K1)/2 differences ?ij ?
• Intervals require sample variance calculation
  based on samples at all K options
• Null hypothesis is accepted if evidence suggests
  all differences ?ij lie in their respective
  intervals
• Null hypothesis is rejected if evidence suggests
  at least one ?ij lies outside its respective
  interval

6
Example Widths of 95 Acceptance Intervals
Increasing with K in TukeyKramer Test
(n1n2nK10)
7
Example of TukeyKramer Test (Example 12.2 in
ISSO)

• Goal With K 4, test null hypothesis L(?1)
  L(?2) L(?3) L(?4) based on 10 measurements at
  each ?i
• All (six) differences ?ij ?
  must lie in acceptance intervals 1.23, 1.23
• Find that ?34 1.72
• Have ?34 ? 1.23, 1.23
• Since at least one ?ij is not in acceptance
  interval, reject null hypothesis
• Conclude at least one ?i likely better than
  others
• Further analysis required to find ?i that is
  better

8
Multiple Comparisons Against One Candidate

• Assume prior information suggests one of K points
  is optimal, say ?m
• Reduces number of comparisons from K(K1)/2
  differences ?ij to only K1
  differences ?mj
• Under null hypothesis, L(?m) ? L(?j) for all j
• Aim to reject null hypothesis
• Implies that L(?m) lt L(?j) for at least some j
• Tests based on critical values lt 0 for
  observed differences ?mj
• To show that L(?m) lt L(?j) for all j requires
  additional analysis

9
Example of Many-to-One Test with Known Variances
(Example 12.3 in ISSO)

• Suppose K 4, m 2 ? Need to compute 3
  critical values , , and for
  acceptance regions
• Valid to take
• Under Bonferroni/Chebyshev
• Under Bonferroni/normal noise
• Under Slepian/normal noise
• Note tighter (smaller) acceptance regions when
  assuming normal noise

10
Widths of 95 Acceptance Intervals (lt 0) for
Tukey-Kramer and Many-to-One Tests (n1n2nK10)
11
Ranking and SelectionIndifference Zone Methods

• Consider usual problem of determining best of K
  possible options, represented ?1 , ?2 ,, ?K
• Have noisy loss measurements yk(?i )
• Suppose analyst is willing to accept any ?i such
  that L(?i) is in indifference zone L(??), L(??)
  ?)
• Analyst can specify ? such that
• P(correct selection of ? ??) ? 1??? ?
• whenever L(?i)??? L(??) ? ? for all ?i ? ??
• Can use independent sampling or common random
  numbers (see Section 14.5 of ISSO)