optiprob

Create a structure describing the optimization problem

Syntax

prob = optiprob('param1',value1,'param2',value2,...)

optiprob

prob = optiprob

prob = optiprob(oldprob,'param1',value1,...)

Description

Note from v1.53 this function is no longer required to build an opti object.

The function optiprob creates the problem structure that you pass to opti for building an optimization problem.

prob = optiprob('param1',value1,'param2',value2,...) creates a OPTI problem structure in which the specified parameters (param) have the specified values (value).

optiprob with no input or output arguments displays a complete list of parameters with their valid names.

prob = optiprob with no input arguments creates a default problem structure

options = optiprob(oldprob,'param1',value1,...) updates the existing object structure, oldprob, with the new values specified.

Fields

The following table lists the available fields for optiprob, noting some field have more than one name associated for ease of use:

Parameter	Value	Description	Default
Name	string	Optional Problem Name	'OPTI Problem'
f / c / grad	1) double column vector or 2) function handle	1) LP / QP linear gradient vector 2) Nonlinear Objective Function Gradient	[]
H / hess	1) double matrix or 2) function handle	1) QP Quadratic Hessian Matrix 2) Nonlinear Hessian of the Lagrangian	[]
Hstr	function handle (sparse)	Nonlinear Hessian Structure	[]
fun / obj	function handle	Nonlinear Objective Function	[]
sense	double scalar	Optimization Sense (1 Minimization, -1 Maximization)	1
objbias	double scalar	Linear or Quadratic Objective Bias Term	0
A	double matrix	Linear Inequality Constraints LHS (Ax <= b) or System of Linear Equations LHS (A)	[]
b	double column vector	Linear Inequality Constraints RHS or System of Linear Equations RHS (b)	[]
Aeq	double matrix	Linear Equality Constraints LHS (Aeqx = beq)	[]
beq	double column vector	Linear Equality Constraints RHS	[]
rl	double column vector	Linear Constraint Lowed Bound (rl <= A*x <= ru)	[]
ru	double column vector	Linear Constraint Upper Bound	[]
lb	double column vector	Lower Bounds (lb <= x <= ub)	[]
ub	double column vector	Upper Bounds	[]
Q	1) double matrix or 2) column cell array of double matrices	1) Quadratic Constraint Quadratic LHS (x'Qx + l'x <= r) 2) A cell array of (1)	[]
l	1) double column vector or 2) double matrix	1) Quadratic Constraint Linear LHS 2) A matrix of (1), each column representing a quadratic constraint l	[]
qrl	1) double scalar or 2) double column vector	1) Quadratic Constraint Lower Bound 2) A vector of (1), each row representing a quadratic constraint qrl	[]
qru	1) double scalar or 2) double column vector	1) Quadratic Constraint Upper Bound 2) A vector of (1), each row representing a quadratic constraint qru	[]
sdcone	1) double matrix 2) column cell array of double matrices	1) Semidefinite constraint (sum(xi*Fi) - F0 >= 0) as [F0(:) F1(:) ... Fn(:)] 2) A cell array of (1)	[]
sedumi	structure	A structure with fields At, b, c, and K representing a SeDuMi problem	[]
nlcon	function handle	Nonlinear Constraint Function	[]
nljac	function handle (sparse)	Nonlinear Constraint Function Jacobian	[]
nljacstr	function handle (sparse)	Nonlinear Constraint Function Jacobian Structure	[]
nlrhs	double column vector	Nonlinear Constraint RHS	[]
nle	double column vector	Nonlinear Constraint Type (-1 <=, 0 ==, 1 >=)	[]
cl	double column vector	Nonlinear Constraint Lower Bound (cl <= nlcon(x) <= cu)	[]
cu	double column vector	Nonlinear Constraint Upper Bound	[]
int / ctype / vtype / ivars	1) character row vector 2) double column vector	1) Integer Variable String (C Continuous, I Integer, B Binary) 2) Integer Variable Indices (e.g. [1 3 5] for 'ICICI')	[]
sos	structure	SOS constraint structure (.type, .index, .weight)	[]
xdata	double column vector	Data Fitting Problem x data	[]
ydata	double column vector	Data Fitting Problem y data	[]
x0 / theta0	double column vector	Initial Solution Guess	[]
ode	function handle	Dynamic Optimization ODE Function	[]
z0 / odez0	double column vector	Dynamic Optimization ODE Initial State Vector	[]

Alternative Entry Groups

To make entering groups of fields easier, several extra arguments are recognized by optiprob for entering common groups:

Group	Parameters	Description	Example
ineq	A, b	Linear Inequality Constraints, A + b	optiprob('ineq',A,b)
eq	Aeq, beq	Linear Equality Constraints, Aeq + beq	optiprob('eq',Aeq,beq)
lin	A, rl, ru	Linear Constraints A + rl + ru	optiprob('lin',A,rl,ru)
bounds	lb, ub	Lower and Upper Bounds	optiprob('bounds',lb,ub)
mix	A, b, e	Mixed Linear Constraints and Type, A + b + e where e is (-1 <=, 0 ==, 1 >=)	optiprob('mix',A,b,e)
ndec	x0 = NaN(ndec,1)	Sets the number of decision variables in a nonlinear problem	optiprob('ndec',2)
nlmix	nlcon, nlrhs, nle	Mixed Nonlinear Constraints and Type, nlcon + nlrhs + nle	optiprob('nlmix',nlcon,nlrhs,nle)
nl	nlcon, cl, cu	Nonlinear Constraints nlcon + cl + cu	optiprob('nl',nlcon,cl,cu)
sle	A, b	System of Linear Equations, A + b	optiprob('sle',A,b)
qp	H, f	Quadratic Objective Function, H + f	optiprob('qp',H,f)
qc	Q, l, qru	Quadratic Inequality Constraints, Q + l + qru	optiprob('qc',Q,l,qru)
qcrow	Q, l, qrl, qru	Quadratic Inequality Constraints, Q + l + qrl + qru	optiprob('qcrow',Q,l,qrl,qru)
sos	sostype, sosind, soswt	Special Ordered Set (SOS) type + indices + weights	optiprob('sos',type,index,weight)
data	xdata, ydata	Data Fitting Problem data	optiprob('data',xdata,ydata)

Examples

Creating a LP

% LP Objective & Constraints
f = -[6 5]';
A = ([1,4; 6,4; 2, -5]);
b = [16;28;6];

% Build Problem
prob = optiprob('grad',f,'ineq',A,b,'bounds',[0;0],[10;10]);

Creating a MILP

% MILP Objective & Constraints
f = -[6 5]';
A = ([1,4; 6,4; 2, -5]);
b = [16;28;6];

% Build Problem
prob = optiprob('grad',f,'ineq',A,b,'bounds',[0;0],[10;10],'int','II');

Creating a NLP

% NLP Objective
obj = @(x) (x(1) - x(2))^2 + (x(2) + x(3) - 2)^2 + (x(4) - 1)^2 + (x(5) - 1)^2;

% NLP Constraints
nlcon = @(x) [ x(1) + 3*x(2);
               x(3) + x(4) - 2*x(5);
               x(2) - x(5) ];
nlrhs = [4 0 0]';
nle   = [0 0 0]';

% Build Problem
prob = optiprob('obj',obj,'nlmix',nlcon,nlrhs,nle,'ndec',5);