ruspy.estimation.est_cost_params.derivative_loglike_cost_params

ruspy.estimation.est_cost_params.derivative_loglike_cost_params(params, maint_func, maint_func_dev, num_states, trans_mat, state_mat, decision_mat, disc_fac, scale, alg_details)

sums up the Jacobian to obtain the gradient of the negative log likelihood function needed for algorithm such as the L-BFGS-B.

Parameters

paramspandas.DataFrame

see Cost parameters

maint_func: func

see Maintenance cost function

maint_func_dev: func

see Maintenance cost function

num_statespython:int

The size of the state space.

trans_matnumpy.array

see Transition matrix

state_matnumpy.array

see State matrix

decision_matnumpy.array

see Decision Matrix

disc_facnumpy.float

see Discount factor

scalenumpy.float

see Scale

alg_detailspython:dict

see :ref: alg_details

Returns

devnp.array

A dimension(params) sized vector containing the gradient of the negative likelihood function.