Optimisation in space of measures and optimal design

Ilya Molchanov; Sergei Zuyev

ESAIM: Probability and Statistics (2010)

  • Volume: 8, page 12-24
  • ISSN: 1292-8100

Abstract

top
The paper develops an approach to optimal design problems based on application of abstract optimisation principles in the space of measures. Various design criteria and constraints, such as bounded density, fixed barycentre, fixed variance, etc. are treated in a unified manner providing a universal variant of the Kiefer-Wolfowitz theorem and giving a full spectrum of optimality criteria for particular cases. Incorporating the optimal design problems into conventional optimisation framework makes it possible to use the whole arsenal of descent algorithms from the general optimisation literature for finding optimal designs. The corresponding steepest descent involves adding a signed measure at every step and converges faster than the conventional sequential algorithms used to construct optimal designs. We study a new class of design problems when the observation points are distributed according to a Poisson point process arising in the situation when the total control on the placement of measurements is impossible.

How to cite

top

Molchanov, Ilya, and Zuyev, Sergei. "Optimisation in space of measures and optimal design." ESAIM: Probability and Statistics 8 (2010): 12-24. <http://eudml.org/doc/104314>.

@article{Molchanov2010,
abstract = { The paper develops an approach to optimal design problems based on application of abstract optimisation principles in the space of measures. Various design criteria and constraints, such as bounded density, fixed barycentre, fixed variance, etc. are treated in a unified manner providing a universal variant of the Kiefer-Wolfowitz theorem and giving a full spectrum of optimality criteria for particular cases. Incorporating the optimal design problems into conventional optimisation framework makes it possible to use the whole arsenal of descent algorithms from the general optimisation literature for finding optimal designs. The corresponding steepest descent involves adding a signed measure at every step and converges faster than the conventional sequential algorithms used to construct optimal designs. We study a new class of design problems when the observation points are distributed according to a Poisson point process arising in the situation when the total control on the placement of measurements is impossible. },
author = {Molchanov, Ilya, Zuyev, Sergei},
journal = {ESAIM: Probability and Statistics},
keywords = {Optimal experimental design; generalized equivalence theorem; constrained optimal design; Poisson design; optimization on measures; gradient methods.; gradient methods},
language = {eng},
month = {3},
pages = {12-24},
publisher = {EDP Sciences},
title = {Optimisation in space of measures and optimal design},
url = {http://eudml.org/doc/104314},
volume = {8},
year = {2010},
}

TY - JOUR
AU - Molchanov, Ilya
AU - Zuyev, Sergei
TI - Optimisation in space of measures and optimal design
JO - ESAIM: Probability and Statistics
DA - 2010/3//
PB - EDP Sciences
VL - 8
SP - 12
EP - 24
AB - The paper develops an approach to optimal design problems based on application of abstract optimisation principles in the space of measures. Various design criteria and constraints, such as bounded density, fixed barycentre, fixed variance, etc. are treated in a unified manner providing a universal variant of the Kiefer-Wolfowitz theorem and giving a full spectrum of optimality criteria for particular cases. Incorporating the optimal design problems into conventional optimisation framework makes it possible to use the whole arsenal of descent algorithms from the general optimisation literature for finding optimal designs. The corresponding steepest descent involves adding a signed measure at every step and converges faster than the conventional sequential algorithms used to construct optimal designs. We study a new class of design problems when the observation points are distributed according to a Poisson point process arising in the situation when the total control on the placement of measurements is impossible.
LA - eng
KW - Optimal experimental design; generalized equivalence theorem; constrained optimal design; Poisson design; optimization on measures; gradient methods.; gradient methods
UR - http://eudml.org/doc/104314
ER -

References

top
  1. A.C. Atkinson and A.N. Donev, Optimum Experimental Designs. Clarendon Press, Oxford (1992).  
  2. C.L. Atwood, Sequences converging to D-optimal designs of experiments. Ann. Statist.1 (1973) 342-352.  
  3. C.L. Atwood, Convergent design sequences, for sufficiently regular optimality criteria. Ann. Statist.4 (1976) 1124-1138.  
  4. D. Böhning, A vertex-exchange-method in D-optimal design theory. Metrika33 (1986) 337-347.  
  5. R. Cominetti, Metric regularity, tangent sets, and second-order optimality conditions. Appl. Math. Optim.21 (1990) 265-287.  
  6. D. Cook and V. Fedorov, Constrained optimization of experimental design. Statistics26 (1995) 129-178.  
  7. D.J. Daley and D. Vere–Jones, An Introduction to the Theory of Point Processes. Springer, New York (1988).  
  8. N. Dunford and J.T. Schwartz, Linear Operators. Part I: General Theory. Wiley, New York (1988).  
  9. V.V. Fedorov, Theory of Optimal Experiments. Academic Press, New York (1972).  
  10. V.V. Fedorov, Optimal design with bounded density: Optimization algorithms of the exchange type. J. Statist. Plan. Inf.22 (1989) 1-13.  
  11. V.V. Fedorov and P. Hackl, Model-Oriented Design of Experiments. Springer, New York, Lecture Notes in Statist. 125 (1997).  
  12. I. Ford, Optimal Static and Sequential Design: A Critical Review, Ph.D. Thesis. Department of Statistics, University of Glasgow, Glasgow (1976).  
  13. A. Gaivoronski, Linearization methods for optimization of functionals which depend on probability measures. Math. Progr. Study28 (1986) 157-181.  
  14. N. Gaffke and R. Mathar, On a Class of Algorithms from Experimental Design Theory. Optimization24 (1992) 91-126.  
  15. E. Hille and R.S. Phillips, Functional Analysis and Semigroups. American Mathematical Society, Providence, AMS Colloquium Publications XXXI (1957).  
  16. J. Kiefer, General equivalence theory for optimum designs (approximate theory). Ann. Statist.2 (1974) 849-879.  
  17. J. Kiefer and J. Wolfowitz, The equivalence of two extremal problems. Canad. J. Math.14 (1960) 363-366.  
  18. P. Kumaravelu, L. Hook, A.M. Morrison, J. Ure, S. Zhao, S. Zuyev, J. Ansell and A. Medvinsky, Quantitative developmental anatomy of definitive haematopoietic stem cells/long-term repopulating units (HSC/RUs): Role of the aorta-gonad-mesonephros (AGM) region and the yolk sac in colonisation of the mouse embryonic liver. Development129 (2002) 4891-4899.  
  19. E.P. Liski, N.K. Mandal, K.R. Shah and B.K. Singha, Topics in Optimal Design. Springer, New York, Lect. Notes Statist. 163 (2002).  
  20. H. Maurer and J. Zowe, First and second-order necessary and sufficient optimality conditions for infinite-dimensional programming problems. Math. Programming16 (1979) 98-110.  
  21. I. Molchanov and S. Zuyev, Steepest descent algorithms in space of measures. Statist. and Comput.12 (2002) 115-123.  
  22. I. Molchanov and S. Zuyev, Tangent sets in the space of measures: With applications to variational calculus. J. Math. Anal. Appl.249 (2000) 539-552.  
  23. I. Molchanov and S. Zuyev, Variational analysis of functionals of a Poisson process. Math. Oper. Res.25 (2000) 485-508.  
  24. C.H. Müller and A. Pázman, Applications of necessary and sufficient conditions for maximin efficient designs. Metrika48 (1998) 1-19.  
  25. A. Pázman, Hilbert-space methods in experimantal design. Kybernetika14 (1978) 73-84.  
  26. E. Polak, Optimization. Algorithms and Consistent Approximations. Springer, New York (1997).  
  27. F. Pukelsheim, Optimal Design of Experiments. Wiley, New York (1993).  
  28. S.M. Robinson, First order conditions for general nonlinear optimization. SIAM J. Appl. Math.30 (1976) 597-607.  
  29. S.D. Silvey, Optimum Design. Chapman & Hall, London (1980).  
  30. D. Stoyan, W.S. Kendall and J. Mecke, Stochastic Geometry and its Applications, Second Edition. Wiley, Chichester (1995).  
  31. P. Whittle, Some general points in the theory of optimal experimental design. J. Roy. Statist. Soc. Ser. B35 (1973) 123-130.  
  32. G. Winkler, Extreme points of moment sets. Math. Oper. Res.13 (1988) 581-587.  
  33. C.-F. Wu, Some algorithmic aspects of the theory of optimal design. Ann. Statist.6 (1978) 1286-1301.  
  34. C.-F. Wu, Some iterative procedures for generating nonsingular optimal designs. Comm. Statist. Theory Methods A7 (1978) 1399-1412.  
  35. C.-F. Wu and H.P. Wynn, The convergence of general step-length algorithms for regular optimum design criteria. Ann. Statist.6 (1978) 1273-1285.  
  36. H.P. Wynn, The sequential generation of D-optimum experimental designs. Ann. Math. Statist.41 (1970) 1655-1664.  
  37. H.P. Wynn, Results in the theory and construction of D-optimum experimental designs. J. Roy. Statist. Soc. Ser. B34 (1972) 133-147.  
  38. J. Zowe and S. Kurcyusz, Regularity and stability for the mathematical programming problem in Banach spaces. Appl. Math. Optim.5 (1979) 49-62.  

NotesEmbed ?

top

You must be logged in to post comments.