horizon.solvers package

Submodules

horizon.solvers.blocksqp module

class horizon.solvers.blocksqp.BlockSqpSolver(prb: horizon.problem.Problem, opts: Dict[KT, VT])

Bases: horizon.solvers.nlpsol.NlpsolSolver

configure_rti() → bool

Take appropriate actions to make the solver suitable for a RTI scheme

Returns:success flag Return type:bool

horizon.solvers.ilqr module

class horizon.solvers.ilqr.SolverILQR(prb: horizon.problem.Problem, opts: Dict[KT, VT] = None)

Bases: horizon.solvers.solver.Solver

configure_rti() → bool

Take appropriate actions to make the solver suitable for a RTI scheme

Returns:success flag Return type:bool

getDt()

Get the horizon’s dt as a vector of dt for each node.

Returns:array of dt values for each node Return type:array

getSolutionDict()

Get the horizon’s solution

Returns:dictionary of variable optimized Return type:dict

print_timings()

save()

set_iteration_callback(cb=None)

solve()

Solve the horizon’s problem

Returns:success flag Return type:bool

horizon.solvers.ipopt module

class horizon.solvers.ipopt.IpoptSolver(prb: horizon.problem.Problem, opts: Dict[KT, VT])

Bases: horizon.solvers.nlpsol.NlpsolSolver

horizon.solvers.nlpsol module

class horizon.solvers.nlpsol.NlpsolSolver(prb: horizon.problem.Problem, opts: Dict[KT, VT], solver_plugin: str)

Bases: horizon.solvers.solver.Solver

build()

fill the dictionary “state_var_impl”

• key: nodes (nNone, n0, n1, …) nNone contains single variables that are not projected in nodes
• val: dict with name and value of implemented variable

getConstraintSolutionDict()

getDt()

Get the horizon’s dt as a vector of dt for each node.

Returns:array of dt values for each node Return type:array

getSolutionDict()

Get the horizon’s solution

Returns:dictionary of variable optimized Return type:dict

solve() → bool

Solve the horizon’s problem

Returns:success flag Return type:bool

horizon.solvers.solver module

class horizon.solvers.solver.Solver(prb: horizon.problem.Problem, opts: Dict[KT, VT] = None)

Bases: abc.ABC

Solver is an abstract interface for a generic solver that aims to find a solution of an instance of horizon.Problem.

The typical usage requires to (i) given an horizon.Problem, call the factory make_solver() to retrieve a solver concrete instance, (ii) to call solve() on it, and (iii) to retrieve the solution as solver.x_opt, and solver.u_opt.

configure_rti() → bool

Take appropriate actions to make the solver suitable for a RTI scheme

Returns:success flag Return type:bool

getDt()

Get the horizon’s dt as a vector of dt for each node.

Returns:array of dt values for each node Return type:array

getSolutionDict()

Get the horizon’s solution

Returns:dictionary of variable optimized Return type:dict

classmethod make_solver(type: str, prb: horizon.problem.Problem, opts: Dict[KT, VT] = None) → horizon.solvers.solver.Solver

Construct a solver from its type name

Parameters:

• type (str) – a string indicating the solver type (e.g., blocksqp, ipopt, ilqr)
• prb (Problem) – the horizon’s problem instance to be solved
• dt (float) – the discretization step, None if not needed
• opts (Dict, optional) – A solver-dependent Dict of options. Defaults to None.

solve() → bool

Solve the horizon’s problem

Returns:success flag Return type:bool

horizon.solvers.sqp module

class horizon.solvers.sqp.GNSQPSolver(prb: horizon.problem.Problem, opts: Dict[KT, VT], qp_solver_plugin: str)

Bases: horizon.solvers.solver.Solver

getAlpha()

getBeta()

getConstraintNormIterations()

getConstraintSolutionDict()

getDt()

Get the horizon’s dt as a vector of dt for each node.

Returns:array of dt values for each node Return type:array

getHessianComputationTime()

getLineSearchComputationTime()

getObjectiveIterations()

getQPComputationTime()

getSolutionDict()

Get the horizon’s solution

Returns:dictionary of variable optimized Return type:dict

setAlphaMin(alpha_min)

setBeta(beta)

set_iteration_callback(cb=None)

solve() → bool

Solve the horizon’s problem

Returns:success flag Return type:bool

Module contents