Skip to main content
Top
Published in: Optimization and Engineering 3/2020

01-06-2020 | Research Article

Regularized stochastic dual dynamic programming for convex nonlinear optimization problems

Authors: Vincent Guigues, Migual A. Lejeune, Wajdi Tekaya

Published in: Optimization and Engineering | Issue 3/2020

Log in

Activate our intelligent search to find suitable subject content or patents.

search-config
loading …

Abstract

We define a regularized variant of the dual dynamic programming algorithm called DDP-REG to solve nonlinear dynamic programming equations. We extend the algorithm to solve nonlinear stochastic dynamic programming equations. The corresponding algorithm, called SDDP-REG, can be seen as an extension of a regularization of the stochastic dual dynamic programming (SDDP) algorithm recently introduced which was studied for linear problems only and with less general prox-centers. We show the convergence of DDP-REG and SDDP-REG. We assess the performance of DDP-REG and SDDP-REG on portfolio models with direct transaction and market impact costs. In particular, we propose a risk-neutral portfolio selection model which can be cast as a multistage stochastic second-order cone program. The formulation is motivated by the impact of market impact costs on large portfolio rebalancing operations. Numerical simulations show that DDP-REG is much quicker than DDP on all problem instances considered (up to 184 times quicker than DDP) and that SDDP-REG is quicker on the instances of portfolio selection problems with market impact costs tested and much faster on the instance of risk-neutral multistage stochastic linear program implemented (8.2 times faster).

Dont have a licence yet? Then find out more about our products and how to get one now:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

  • über 102.000 Bücher
  • über 537 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Maschinenbau + Werkstoffe
  • Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 390 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Maschinenbau + Werkstoffe




 

Jetzt Wissensvorsprung sichern!

Appendix
Available only for authorised users
Footnotes
1
Note that to simplify notation, the same notation \(\xi _\mathtt{{Index}}\) is used to denote the realization of the process at node Index of the scenario tree and the value of the process \((\xi _t)\) for stage Index. The context will allow us to know which concept is being referred to. In particular, letters n and m will only be used to refer to nodes while t will be used to refer to stages.
 
2
Note that the proposition can be applied because Assumption (H2) holds and thus the assumptions of the proposition are satisfied for value function \({\underline{{\mathfrak {Q}}}}_t^k( \cdot , \xi _m )\).
 
3
In Guigues (2016) a forward, instead of a forward-backward algorithm, is considered. In this setting, finiteness of coefficients \(\theta _t^k\) and \(\beta _t^k\) is not guaranteed for the first iterations (for instance \((\theta _t^1)_t\) are \(-\infty\)) but the proof is similar.
 
4
It is indeed immediately seen that (4.3) and (4.4) is of form (2.2)–(2.4), writing the maximization problems as minimization problems and introducing the extended state \(s_t=(x_t, y_t, z_t)\).
 
5
Though when deriving these relations in (i) we had fixed \(k \in {\mathcal {S}}_n\), the inequalities we now re-use for (ii) are valid for any \(k \ge 1\).
 
6
The existence of an accumulation point comes from the fact that the decisions belong almost surely to a compact set.
 
Literature
go back to reference Almgren R (2003) Optimal execution with nonlinear impact functions and trading-enhanced risk. Appl Math Finance 10:1–18CrossRef Almgren R (2003) Optimal execution with nonlinear impact functions and trading-enhanced risk. Appl Math Finance 10:1–18CrossRef
go back to reference Almgren R, Thum C, Li H (2005) Equity market impact. Risk 18:57–62 Almgren R, Thum C, Li H (2005) Equity market impact. Risk 18:57–62
go back to reference Asamov T, Powell W (2015) Regularized decomposition of high-dimensional multistage stochastic programs with Markov uncertainty. SIAM J Optim 28:575–595MathSciNetCrossRef Asamov T, Powell W (2015) Regularized decomposition of high-dimensional multistage stochastic programs with Markov uncertainty. SIAM J Optim 28:575–595MathSciNetCrossRef
go back to reference Bandarra M, Guigues V (2019) Single cut and multicut SDDP with cut selection for multistage stochastic linear programs: convergence proof and numerical experiments. arXiv:1902.06757 Bandarra M, Guigues V (2019) Single cut and multicut SDDP with cut selection for multistage stochastic linear programs: convergence proof and numerical experiments. arXiv:​1902.​06757
go back to reference Bouchaud J, Gefen Y, Potters M, Wyart M (2004) Fluctuations and response in financial markets: the subtle nature ofrandom price changes. Quant Finance 4:176–190CrossRef Bouchaud J, Gefen Y, Potters M, Wyart M (2004) Fluctuations and response in financial markets: the subtle nature ofrandom price changes. Quant Finance 4:176–190CrossRef
go back to reference Cadenillas A (2000) Consumption-investment problems with transaction costs: survey and open problems. Math Methods Oper Res 51:43–68MathSciNetCrossRef Cadenillas A (2000) Consumption-investment problems with transaction costs: survey and open problems. Math Methods Oper Res 51:43–68MathSciNetCrossRef
go back to reference de Matos V, Philpott A, Finardi E (2015) Improving the performance of stochastic dual dynamic programming. J Comput Appl Math 290:196–208MathSciNetCrossRef de Matos V, Philpott A, Finardi E (2015) Improving the performance of stochastic dual dynamic programming. J Comput Appl Math 290:196–208MathSciNetCrossRef
go back to reference Filomena T, Lejeune M (2012) Stochastic portfolio optimization with proportional transaction costs: convex reformulations and computational experiments. Oper Res Lett 40:212–217MathSciNetCrossRef Filomena T, Lejeune M (2012) Stochastic portfolio optimization with proportional transaction costs: convex reformulations and computational experiments. Oper Res Lett 40:212–217MathSciNetCrossRef
go back to reference Frino A, Bjursell J, Wang G, Lepone A (2008) Large trades and intraday futures price behavior. J Fut Mark 28:1117–1181CrossRef Frino A, Bjursell J, Wang G, Lepone A (2008) Large trades and intraday futures price behavior. J Fut Mark 28:1117–1181CrossRef
go back to reference Gabaix X, Gopikrishnan P, Plerou V, Stanley H (2003) A theory of power-law distributions in financial market fluctuations. Nature 423:267–270CrossRef Gabaix X, Gopikrishnan P, Plerou V, Stanley H (2003) A theory of power-law distributions in financial market fluctuations. Nature 423:267–270CrossRef
go back to reference Girardeau P, Leclere V, Philpott A (2015) On the convergence of decomposition methods for multistage stochastic convex programs. Math Oper Res 40:130–145MathSciNetCrossRef Girardeau P, Leclere V, Philpott A (2015) On the convergence of decomposition methods for multistage stochastic convex programs. Math Oper Res 40:130–145MathSciNetCrossRef
go back to reference Grinold R (2006) A dynamic model of portfolio management. J Invest Manag 4:5–22 Grinold R (2006) A dynamic model of portfolio management. J Invest Manag 4:5–22
go back to reference Grinold R, Kahn R (2000) Active Portfolio Management, 2nd edn. McGraw-Hill, New York Grinold R, Kahn R (2000) Active Portfolio Management, 2nd edn. McGraw-Hill, New York
go back to reference Guigues V (2014) SDDP for some interstage dependent risk-averse problems and application to hydro-thermal planning. Comput Optim Appl 57:167–203MathSciNetCrossRef Guigues V (2014) SDDP for some interstage dependent risk-averse problems and application to hydro-thermal planning. Comput Optim Appl 57:167–203MathSciNetCrossRef
go back to reference Guigues V (2016) Convergence analysis of sampling-based decomposition methods for risk-averse multistage stochastic convex programs. SIAM J Optim 26:2468–2494MathSciNetCrossRef Guigues V (2016) Convergence analysis of sampling-based decomposition methods for risk-averse multistage stochastic convex programs. SIAM J Optim 26:2468–2494MathSciNetCrossRef
go back to reference Guigues V (2017) Dual dynamic programing with cut selection: convergence proof and numerical experiments. Eur J Oper Res 258:47–57MathSciNetCrossRef Guigues V (2017) Dual dynamic programing with cut selection: convergence proof and numerical experiments. Eur J Oper Res 258:47–57MathSciNetCrossRef
go back to reference Guigues V, Römisch W (2012a) Sampling-based decomposition methods for multistage stochastic programs based on extended polyhedral risk measures. SIAM J Optim 22:286–312MathSciNetCrossRef Guigues V, Römisch W (2012a) Sampling-based decomposition methods for multistage stochastic programs based on extended polyhedral risk measures. SIAM J Optim 22:286–312MathSciNetCrossRef
go back to reference Guigues V, Römisch W (2012b) SDDP for multistage stochastic linear programs based on spectral risk measures. Oper Res Lett 40:313–318MathSciNetCrossRef Guigues V, Römisch W (2012b) SDDP for multistage stochastic linear programs based on spectral risk measures. Oper Res Lett 40:313–318MathSciNetCrossRef
go back to reference Infanger G, Morton D (1996) Cut sharing for multistage stochastic linear programs with interstage dependency. Math Program 75:241–256MathSciNetMATH Infanger G, Morton D (1996) Cut sharing for multistage stochastic linear programs with interstage dependency. Math Program 75:241–256MathSciNetMATH
go back to reference Kozmik V, Morton D (2015) Evaluating policies in risk-averse multi-stage stochastic programming. Math Program 152:275–300MathSciNetCrossRef Kozmik V, Morton D (2015) Evaluating policies in risk-averse multi-stage stochastic programming. Math Program 152:275–300MathSciNetCrossRef
go back to reference Lemarechal C (1974) An algorithm for minimizing convex functions. In: Proceedings of the IFIP’74, Stockholm Lemarechal C (1974) An algorithm for minimizing convex functions. In: Proceedings of the IFIP’74, Stockholm
go back to reference Lillo F, Farmer J, Mantegna R (2003) Econophysics: master curve for price-impact function. Nature 421:129–130CrossRef Lillo F, Farmer J, Mantegna R (2003) Econophysics: master curve for price-impact function. Nature 421:129–130CrossRef
go back to reference Loeb T (1983) Trading costs: the critical link between investment information and results. Financ Anal J 39:39–44CrossRef Loeb T (1983) Trading costs: the critical link between investment information and results. Financ Anal J 39:39–44CrossRef
go back to reference Mitchell J, Braun S (2013) Rebalancing an investment portfolio in the presence of convex transaction costs, including market impact costs. Optim Methods Softw 28:523–542MathSciNetCrossRef Mitchell J, Braun S (2013) Rebalancing an investment portfolio in the presence of convex transaction costs, including market impact costs. Optim Methods Softw 28:523–542MathSciNetCrossRef
go back to reference Mo B, Gjelsvik A, Grundt A (2001) Integrated risk management of hydro power scheduling and contract management. IEEE Trans Power Syst 16:216–221CrossRef Mo B, Gjelsvik A, Grundt A (2001) Integrated risk management of hydro power scheduling and contract management. IEEE Trans Power Syst 16:216–221CrossRef
go back to reference Moazeni S, Coleman T, Li Y (2010) Optimal portfolio execution strategies and sensitivity to price impact parameters. SIAM J Optim 20:1620–1654MathSciNetCrossRef Moazeni S, Coleman T, Li Y (2010) Optimal portfolio execution strategies and sensitivity to price impact parameters. SIAM J Optim 20:1620–1654MathSciNetCrossRef
go back to reference Moro E, Vicente J, Moyano L, Gerig A, Farmer J, Vaglica G, Lillo F, Mantegna R (2009) Market impact and trading profile of hidden orders in stock markets. Phys Rev E 80:1–8CrossRef Moro E, Vicente J, Moyano L, Gerig A, Farmer J, Vaglica G, Lillo F, Mantegna R (2009) Market impact and trading profile of hidden orders in stock markets. Phys Rev E 80:1–8CrossRef
go back to reference MOSEK (2017) MOSEK optimization suite. release 8.0.0.52 MOSEK (2017) MOSEK optimization suite. release 8.0.0.52
go back to reference Pereira M, Pinto L (1991) Multi-stage stochastic optimization applied to energy planning. Math Program 52:359–375MathSciNetCrossRef Pereira M, Pinto L (1991) Multi-stage stochastic optimization applied to energy planning. Math Program 52:359–375MathSciNetCrossRef
go back to reference Pfeiffer L, Apparigliato R, Auchapt S (2012) Two methods of pruning benders’ cuts and their application to the management of a gas portfolio. Research report RR-8133, hal-00753578 Pfeiffer L, Apparigliato R, Auchapt S (2012) Two methods of pruning benders’ cuts and their application to the management of a gas portfolio. Research report RR-8133, hal-00753578
go back to reference Philpott AB, de Matos V (2012) Dynamic sampling algorithms for multi-stage stochastic programs with risk aversion. Eur J Oper Res 218:470–483MathSciNetCrossRef Philpott AB, de Matos V (2012) Dynamic sampling algorithms for multi-stage stochastic programs with risk aversion. Eur J Oper Res 218:470–483MathSciNetCrossRef
go back to reference Philpott AB, Guan Z (2008) On the convergence of stochastic dual dynamic programming and related methods. Oper Res Lett 36:450–455MathSciNetCrossRef Philpott AB, Guan Z (2008) On the convergence of stochastic dual dynamic programming and related methods. Oper Res Lett 36:450–455MathSciNetCrossRef
go back to reference Rockafellar R, Uryasev S (2002) Conditional value-at-risk for general loss distributions. J Bank Finance 26:1443–1471CrossRef Rockafellar R, Uryasev S (2002) Conditional value-at-risk for general loss distributions. J Bank Finance 26:1443–1471CrossRef
go back to reference Sen S, Zhou Z (2014) Multistage stochastic decomposition: a bridge between stochastic programming and approximate dynamic programming. SIAM J Optim 24:127–153MathSciNetCrossRef Sen S, Zhou Z (2014) Multistage stochastic decomposition: a bridge between stochastic programming and approximate dynamic programming. SIAM J Optim 24:127–153MathSciNetCrossRef
go back to reference Shapiro A, Dentcheva D, Ruszczyński A (2009) Lectures on stochastic programming: modeling and theory. SIAM, PhiladelphiaCrossRef Shapiro A, Dentcheva D, Ruszczyński A (2009) Lectures on stochastic programming: modeling and theory. SIAM, PhiladelphiaCrossRef
go back to reference Shapiro A, Tekaya W, da Costa J, Soares M (2013) Risk neutral and risk averse stochastic dual dynamic programming method. Eur J Oper Res 224:375–391MathSciNetCrossRef Shapiro A, Tekaya W, da Costa J, Soares M (2013) Risk neutral and risk averse stochastic dual dynamic programming method. Eur J Oper Res 224:375–391MathSciNetCrossRef
go back to reference Tikhonov A (1943) On the stability of inverse problems. Dokl Akad Nauk SSSR 39:195–198MathSciNet Tikhonov A (1943) On the stability of inverse problems. Dokl Akad Nauk SSSR 39:195–198MathSciNet
go back to reference Torre N (1997) Market impact model handbook. BARRA Inc., Berkeley Torre N (1997) Market impact model handbook. BARRA Inc., Berkeley
go back to reference Zagst R, Kalin D (2007) Portfolio optimization under liquidity costs. Int J Pure Appl Math 39:217–233MathSciNetMATH Zagst R, Kalin D (2007) Portfolio optimization under liquidity costs. Int J Pure Appl Math 39:217–233MathSciNetMATH
Metadata
Title
Regularized stochastic dual dynamic programming for convex nonlinear optimization problems
Authors
Vincent Guigues
Migual A. Lejeune
Wajdi Tekaya
Publication date
01-06-2020
Publisher
Springer US
Published in
Optimization and Engineering / Issue 3/2020
Print ISSN: 1389-4420
Electronic ISSN: 1573-2924
DOI
https://doi.org/10.1007/s11081-020-09511-0

Other articles of this Issue 3/2020

Optimization and Engineering 3/2020 Go to the issue

Premium Partners