LineSearch

TwoStageOptimalControl.LineSearch — Function

LineSearch(Con::Array{Float64,3}, 
           Stat::Array{Float64,3}, 
           Con_dist::Array{Float64,3},  
           Stat_dist::Array{Float64,3}, 
           Stat_agg::Array{Float64,3}, 
           Dir_u::Array{Float64,3}, 
           Dir_u_dist::Array{Float64,3}, 
           ObjValue::Float64, 
           Step, 
           Para::Dict,
           Con_new::Array{Float64,3}, 
           Stat_new::Array{Float64,3}, 
           Con_dist_new::Array{Float64,3}, 
           Stat_dist_new::Array{Float64,3}, 
           Stat_agg_new::Array{Float64,3},
           Con_best::Array{Float64,3}, 
           Stat_best::Array{Float64,3}, 
           Con_dist_best::Array{Float64,3}, 
           Stat_dist_best::Array{Float64,3}, 
           Stat_agg_best::Array{Float64,3})

Searches for an improvement along the gradient using a quadratic approximation for the objective function along the gradient.

source

TwoStageOptimalControl.Update — Function

Update(Con::Array{Float64,3},
       Con_dist::Array{Float64,3},
       Dir_u::Array{Float64,3},
       Dir_u_dist::Array{Float64,3},
       Step,
       Para::Dict,
       Con_new::Array{Float64,3},
       Stat_new::Array{Float64,3},
       Con_dist_new::Array{Float64,3},
       Stat_dist_new::Array{Float64,3}, 
       Stat_agg_new::Array{Float64,3})

Update calculates the new values for Controls Con_new, States Stat_new and Objectiv value given start values for the Controls Con, the search direction Dir_u and the step-size in the direction step

source

TwoStageOptimalControl.AssignBest — Function

AssignBest(ObjValue::Float64,
           Con::Array{Float64,3},
           Stat::Array{Float64,3},
           Con_dist::Array{Float64,3},
           Stat_dist::Array{Float64,3}, 
           Stat_agg::Array{Float64,3}, 
           step, 
           Con_best::Array{Float64,3}, 
           Stat_best::Array{Float64,3}, 
           Con_dist_best::Array{Float64,3},  
           Stat_dist_best::Array{Float64,3},  
           Stat_agg_best::Array{Float64,3})

Assign the current variables to the "best" variables

source

TwoStageOptimalControl.maxabsGradient — Function

maxabsGradient(Con::Array{Float64,3},
               Con_dist::Array{Float64,3},
               dHam::Array{Float64,3},
               dHam_dist::Array{Float64,3},
               Para::Dict)

Calculate the maximum of the absolute value of the gradient of the Hamiltonian accounting for controls which are on the boundary of their feasible regions.

source