1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
|
// Copyright (C) 2021 EDF
// All Rights Reserved
// This code is published under the GNU Lesser General Public License (GNU LGPL)
#ifndef OPTIMIZERSWITCHBASE_H
#define OPTIMIZERSWITCHBASE_H
#include <Eigen/Dense>
#include "StOpt/core/utils/types.h"
#include "StOpt/core/utils/StateWithIntState.h"
#include "StOpt/regression/BaseRegression.h"
#include "StOpt/dp/SimulatorDPBase.h"
#include "StOpt/core/grids/RegularSpaceIntGrid.h"
/** \file OptimizerSwitchingBase.h
* \brief Define an abstract class for Switching problems solved by regression methods and with an integer state
* \author Xavier Warin
*/
namespace StOpt
{
/// \class OptimizerSwitchBase OptimizerSwitchBase.h
/// Base class for optimizer for Dynamic Programming with regression methods
class OptimizerSwitchBase
{
public :
OptimizerSwitchBase() {}
virtual ~OptimizerSwitchBase() {}
/// \brief defines the diffusion cone for parallelism
/// \param p_iRegime regime used
/// \param p_regionByProcessor region (min max) treated by the processor for the different regimes treated
/// \return returns in each dimension the min max values in the stock that can be reached from the grid p_gridByProcessor for each regime
virtual std::vector< std::array<int, 2 > > getCone(const int &p_iReg, const std::vector< std::array<int, 2 > > &p_regionByProcessor) const = 0;
/// \brief defines the dimension to split for MPI parallelism
/// For each regime : for each dimension return true is the direction can be split
virtual std::vector< Eigen::Array< bool, Eigen::Dynamic, 1> > getDimensionToSplit() const = 0 ;
/// \brief defines a step in optimization
/// \param p_grid grid at arrival step after command (integer states) for each regime
/// \param p_iReg regime treated
/// \param p_state coordinates of the deterministic integer state
/// \param p_condExp Conditional expectation operator
/// \param p_phiIn for each regime gives the solution calculated at the previous step ( next time step by Dynamic Programming resolution) : structure of the 2D array ( nb simulation ,nb stocks )
/// \return a pair :
/// - for each regimes (column) gives the solution for each particle (row)
/// - for each control (column) gives the optimal control for each particle (rows)
/// .
virtual Eigen::ArrayXd stepOptimize(const std::vector<std::shared_ptr< RegularSpaceIntGrid> > &p_grid,
const int &p_iReg,
const Eigen::ArrayXi &p_state,
const std::shared_ptr< BaseRegression> &p_condExp,
const std::vector < std::shared_ptr< Eigen::ArrayXXd > > &p_phiIn) const = 0;
/// \brief defines a step in simulation
/// Notice that this implementation is not optimal but is convenient if the control is discrete.
/// By avoiding interpolation in control we avoid non admissible control
/// Control are recalculated during simulation.
/// \param p_grid grid at arrival step after command for each regime
/// \param p_condExp Conditional expectation operator reconstructing conditionnal expectation from basis functions for each state
/// \param p_basisFunc Basis functions par each point of the grid state for each regime
/// \param p_state defines the state value (modified)
/// \param p_phiInOut defines the value functions (modified) : size number of functions to follow
virtual void stepSimulate(const std::vector< std::shared_ptr< RegularSpaceIntGrid> > &p_grid,
const std::shared_ptr< BaseRegression> &p_condExp,
const std::vector< Eigen::ArrayXXd > &p_basisFunc,
StOpt::StateWithIntState &p_state,
Eigen::Ref<Eigen::ArrayXd> p_phiInOut) const = 0 ;
/// \brief Get the number of regimes allowed for the asset to be reached at the current time step
/// If \f$ t \f$ is the current time, and $\f$ dt \f$ the resolution step, this is the number of regime allowed on \f$[ t- dt, t[\f$
virtual int getNbRegime() const = 0 ;
/// \brief get the simulator back
virtual std::shared_ptr< StOpt::SimulatorDPBase > getSimulator() const = 0;
/// \brief get size of the function to follow in simulation
virtual int getSimuFuncSize() const = 0;
};
}
#endif /* OPTIMIZERSWITCHBASE_H */
|
