[ Search |  Up Level |  Project home |  Index |  Class hierarchy ]

 Class hierarchy

 
Algorithm
 
Base class for optimization and system-solving algorithms.
 
 
GradientBased
 
Gradient based algorithms.
 
 
 
HillClimbing
 
Algorithms that optimize an objective based on gradient and/or Hessian information.
 
 
 
 
Newton
 
Newton Updating of H .
 
 
 
 
 
BHHH
 
Berndt Hall Hall Hausman Updating.
 
 
 
 
QuasiNewton
 
Container for algorithms that use but do not compute the Hessian matrix H.
 
 
 
 
 
BFGS
 
Broyden Fletcher Goldfarb Shanno Updating of the hessian H.
 
 
 
 
 
DFP
 
[not coded yet] Davidon Fletcher Powell Updating of H.
 
 
 
RootFinding
 
Solve system of equations using Jacobian information.
 
 
 
 
Broyden
 
Broyden approximation to the Jacobian.
 
 
 
 
NewtonRaphson
 
Update the Jacobian on each iteration.
 
 
 
SQP
 
Sequential Quadratic Programming for constrained optimization.
 
 
 
 
SQPBFGS
 
 
 
 
 
SQPNEWTON
 
 
 
NonGradient
 
Container for algorithms that do not rely on gradients.
 
 
 
LineMethod
 
Methods specific to solving or optimizing in one dimension.
 
 
 
 
LineMax
 
One-dimensional line search for a maximum.
 
 
 
 
 
CLineMax
 
Constrained line maximization.
 
 
 
 
SysMax
 
Systems line maximization.
 
 
 
 
 
OneDimRoot
 
Solve for the root of a OneDimSystem system using Bracket-Bisect.
 
 
 
NelderMead
 
The Nelder and Mead Amoeba (Simplex) algorithm.
 
 
 
SimulatedAnnealing
 
Metropolis Simulated Annealing algorithm .
 
   
 
RandomSearch
 
A special case of annealing in which the temperature stays the same and only improvements are accepted.
 
ObjClient
 
Client for parallel evaluation of objectives.
 
Objective
 
Base class for objective optimization and system solving.
 
 
Constrained
 
Container for Constrained Objectives.
 
 
System
 
A non-linear system of equations to solve.
 
 
 
Equilibrium
 
A nonlinear system for computing equilibrium.
 
 
 
OneDimSystem
 
A One Dimensional Non-linear system (can be solved with OneDimSolve).
 
 
UnConstrained
 
Container for Unconstrained Objectives.
 
 
 
BlackBox
 
Represents a blacbox objective.
 
 
 
 
CES
 
Constant elasticity of subsitution function.
 
 
 
 
CobbDouglas
 
Cobb-Douglas objective.
 
 
 
 
DataObjective
 
Base class for automatically generated econometric objectives: likelihood functions and GMM objective.
 
 
 
 
MultiNomialChoice
 
Objective for multinomial choice models.
 
 
 
 
 
MLogit
 
Multinomial Logit Model.
 
 
 
 
 
MNP
 
Container for normally-distributed multinomial discrete choice reduced-form models.
 
 
 
 
 
 
GHKMNP
 
Correlated MNP using GHK .
 
 
 
 
 
 
GQMNP
 
Independent MNP using Gausss-Hermite integration.
 
 
 
 
NoObjective
 
 
 
 
Separable
 
Represent sum of K BlackBox objectives.
 
   
 
Mixture
 
 
ObjServer
 
Server for parallel evaluation of BlackBox objectives.
 
 
CstrServer
 
 
 
SepServer
 
 
Quantity
 
Container for discrete variables (DDP) and continuous parameters (FiveO).
 
 
Discrete
 
Discrete values: actions, states, auxiliaries.
 
 
Parameter
 
Continuously varying quantity: the base class for parameters of an Objective.
 
 
 
Determined
 
Value determined exactly by some other value, not chosen by optimization.
 
 
 
Free
 
Can take on any real number: \(-\infty \lt v \lt \infty\).
 
 
 
Limited
 
Container for parameters are neither Free nor Determined.
 
 
 
 
Bounded
 
A parameter contained in an open interval \(L \lt v \lt U\).
 
 
 
 
 
Correlation
 
Bounded as: -1 < v < 1 .
 
 
 
 
 
Probability
 
Bounded as: 0 < v < 1 .
 
 
 
 
BoundedAbove
 
Parameter bounded from above \(-\infty \lt v \lt U\).
 
 
 
 
 
Negative
 
Bounded from above by 0.
 
 
 
 
BoundedBelow
 
A parameter bounded from below: \(L \lt v \lt \infty\).
 
 
 
 
 
Positive
 
Parameter Bounded from below by 0.
 
 
 
ParameterBlock
 
Two or more parameters whose ranges interact or are related for some other reason.
 
   
 
Coefficients
 
Vector of free parameters.
 
   
 
DecreasingReturns
 
Vector of J probabilities that sum to strictly less than 1.
 
   
 
FixedBlock
 
Vector of values determined exactly by some other value.
 
   
 
Ordered
 
 
   
 
 
Decreasing
 
Vector of parameters that are sequentially decreasing.
 
   
 
 
Increasing
 
Vector of parameters that are sequentially increasing.
 
   
 
Probabilities
 
Vector of Probabilities.
 
   
 
Simplex
 
Vector of J probabilities that sum to 1.
 
   
 
StDeviations
 
Positive Vector.
 
Zauxiliary
 
A container for auxiliary structures, which helps organize the hierarchy of classes.
 
 
CGI
 
.
 
 
Discretized
 
A continuous discretization of a (potentially) continuous quantity.
 
 
Equations
 
Code a system of constraints.
 
 
 
Equality
 
A container for equality constraints.
 
 
 
InEquality
 
A container for inequality constraints.
 
 
Integration
 
Container for different integration techniques.
 
 
 
GaussianQuadrature
 
 
 
 
 
GQH
 
Gauss-Hermite Quadrature Integration.
 
 
 
 
GQL
 
Gauss-Laguerre Quadrature Integration.
 
 
 
GHK
 
Smooth Simulation of Multinomial Normal Probabilities and EValues.
 
 
LinePoint
 
Holds one line maximization try.
 
 
 
Point
 
Store information about a multidimensional point.
 
 
 
 
CPoint
 
Store a point for a constrained objective.
 
 
 
 
MixPoint
 
Store a point for a mixture objective.
 
 
 
 
SepPoint
 
Store a point for a separable objective.
 
 
 
 
SysPoint
 
A system point.
 
 
Version
 
Checks minimum Ox version and prints copyright info.