分享我的协同优化的成果——HeartDiople和海沟问题

rabitzh · 发表于 2006-5-30 19:45:04

HeartDIople问题的迭代次数很多，得到的解也不是那么理想，可能是初始值的原因。
其中第一组数据解是：
Execution Results Summary

Start time: Mon May 29 21:15:25 WST 2006
End time: Mon May 29 22:25:56 WST 2006

Task: HeartDiopleCO
  Total runs:    99
  Feasible runs: 0
  Infeasible runs: 99
  Failed runs:    0
  Database file: HeartDiopleCO.db

  Optimization Plan: NewPlan
Executed between RunCounter 1 and 99 (99 runs)

Techniques used:
      Step1: Sequential Quadratic Programming - DONLP

Best design:             currently             previously
--------------------------------------------------------------------
      RunCounter          73                   1
      ObjectiveAndPenalty 4495.16432085206    48369682.8357491
      Objective          48.525598989361       2625.75
      Penalty             4446.6387218627       48367057.0857491

Best design ObjectiveAndPenalty value improved by
48365187.67 (99.99%) after executing this Optimization Plan

Best design parameter values:

      z6 = 1.87059601535712
      z7 = -1.58602535552925
      z8 = -1.44375504429812
      z1 = 8.44674965102361
      z2 = -9.05310286535493
      z3 = -0.484396997889745
      z4 = 3.06553247208648
      z5 = 1.64978851836841
      J2 = 0.51435308963862 (violates UpperBound=0.01)
      J1 = 0.129861198798801 (violates UpperBound=0.01)
      f6 = 14.9323407457401
      f7 = 5.30202171054587
      minF = 48.525598989361
      f8 = 30.332780060864
      f5 = -2.04154352778901  (violates LowerBound=0.0）
第二组为：
Execution Results Summary

Start time: Tue May 30 08:53:20 WST 2006
End time: Tue May 30 14:02:47 WST 2006

Task: HeartDiopleCO
  Total runs:    505
  Feasible runs: 0
  Infeasible runs: 505
  Failed runs:    0
  Database file: HeartDiopleCO.db

  Optimization Plan: NewPlan
Executed between RunCounter 1 and 505 (505 runs)

Techniques used:
      Step1: Sequential Quadratic Programming - DONLP

Best design:             currently             previously
--------------------------------------------------------------------
      RunCounter          500                   1
      ObjectiveAndPenalty 55.0088823824952    98255722282.08
      Objective          -0.0127854814708845 -7670.0
      Penalty             55.0216678639661    98255729952.08

Best design ObjectiveAndPenalty value improved by
9.825572223e+010 (100.0%) after executing this Optimization Plan

Best design parameter values:

      z6 = 0.66769786750296
      z7 = -0.251870457677615
      z8 = 0.35181940505958
      z1 = 6.99997489586061
      z2 = -7.97891666262143
      z3 = 1.16246305596222
      z4 = 0.782198682612054
      z5 = 1.05256347680304
      J2 = 0.0145301822232409  (violates UpperBound=0.01)
      J1 = 0.221339124462653 (violates UpperBound=0.01)
      f6 = 0.00587646539609032
      f7 = -0.0156707222170347 (violates LowerBound=0.0)
      minF = -0.0127854814708845
      f8 = 0.00656640317008339
      f5 = -0.00955762782002356 (violates LowerBound=0.0)
源程序为：
MDOLVersion: 8.0
CompilerOptions: warn

Task HeartDiopleCO

TaskHeader HeartDiopleCO
      Version: 1.0
      Evaluation: taskplan
      ControlMode: user
      RunCounter: 1
      BoundsPolicy: adjustvalue
      CheckPoint: unknown
End TaskHeader HeartDiopleCO

Inputs HeartDiopleCO
      Parameter: z6 Type: real InitialValue: -10.0
      Parameter: z7 Type: real InitialValue: 2.0
      Parameter: z8 Type: real InitialValue: 3.0
      Parameter: z1 Type: real InitialValue: 8.0
      Parameter: z2 Type: real InitialValue: -8.0
      Parameter: z3 Type: real InitialValue: 1.0
      Parameter: z4 Type: real InitialValue: 5.0
      Parameter: z5 Type: real InitialValue: 4.0
End Inputs HeartDiopleCO

Outputs HeartDiopleCO
      Parameter: J2 Type: real
      Parameter: J1 Type: real
      Parameter: f6 Type: real
      Parameter: f7 Type: real
      Parameter: minF Type: real
      Parameter: f8 Type: real
      Parameter: f5 Type: real
End Outputs HeartDiopleCO

Calculations HeartDiopleCO
      Calculation Calculation5
         Parameters
            z6 z7 z8 z1 z2 z3 z4 z5 f6 f7
            minF f8 f5
         Statements
            f5 = z1*(z5^2-z7^2)-2*z3*z5*z7+z2*(z6^2-z8^2)-2*z4*z6*z8-5.0
               f6 = z3*(z5^2-z7^2)+2*z1*z5*z7+z4*(z6^2-z8^2)+2*z2*z6*z8+6.0

               f7=z1*z5*(z5^2-3*z7^2)+z3*z7*(z7^2-3*z5^2)+z2*z6*(z6^2-3*z8^2)+z4*z8*(z8^2-3*z6^2)-7
               f8=z3*z5*(z5^2-3*z7^2)-z1*z7*(z7^2-3*z5^2)+z4*z6*(z6^2-3*z8^2)-z2*z8*(z8^2-3*z6^2)+8

            minF=f5+f6+f7+f8
         End Statements
      End Calculation Calculation5
End Calculations HeartDiopleCO

Task output28

      TaskHeader output28
         Version: 1.0
         Evaluation: taskplan
         ControlMode: user
         RunCounter: 1
         BoundsPolicy: adjustvalue
         CheckPoint: unknown
      End TaskHeader output28

      Inputs output28
         Parameter: z6 Type: real InitialValue: 1.8705959911958
         Parameter: z7 Type: real InitialValue: -1.58718411550964
         Parameter: z8 Type: real InitialValue: -1.44375507956317
         Parameter: x1 Type: real InitialValue: 5.0
         Parameter: x3 Type: real InitialValue: 2.0
         Parameter: z1 Type: real InitialValue: 8.44674964678718
         Parameter: x4 Type: real InitialValue: 1.0
         Parameter: z2 Type: real InitialValue: -9.05310287720698
         Parameter: x5 Type: real InitialValue: -1.0
         Parameter: z3 Type: real InitialValue: -0.484397005726966
         Parameter: x6 Type: real InitialValue: 2.0
         Parameter: z4 Type: real InitialValue: 3.06553246855229
         Parameter: x7 Type: real InitialValue: -3.0
         Parameter: z5 Type: real InitialValue: 1.64978849482124
      End Inputs output28

      Outputs output28
         Parameter: x8 Type: real
         Parameter: J1 Type: real
         Parameter: x2 Type: real
      End Outputs output28

      Calculations output28
         Calculation Calculation3
            Parameters
                  z6 z7 z8 x1 x3 z1 x4 z2 x5 z3
                  x6 z4 x7 z5 x8 x2 J1
            Statements
                  x2 = -1-x1
                  x8=(x5*x1+x6*x2-x7*x3+3)/x4

                  J1=(x1-z1)^2+(x2-z2)^2+(x3-z3)^2+(x4-z4)^2+(x5-z5)^2+(x6-z6)^2+(x7-z7)^2+(x8-z8)^2
            End Statements
         End Calculation Calculation3
      End Calculations output28

      TaskProcess output28
         Control: [
            Calculation3
         ]
      End TaskProcess output28

      Optimization output28
         PotentialVariables:
            z6 z7 z8 x1 x3 z1 x4 z2 x5 z3
            x6 z4 x7 z5
         Variables:
            x1 x3 x4 x5 x6 x7
         VariableScaling
            Parameter: z6 ScaleFactor: 1.0
            Parameter: z7 ScaleFactor: 1.0
            Parameter: z8 ScaleFactor: 1.0
            Parameter: x1 ScaleFactor: 10.0
            Parameter: x3 ScaleFactor: 10.0
            Parameter: z1 ScaleFactor: 1.0
            Parameter: x4 ScaleFactor: 10.0
            Parameter: z2 ScaleFactor: 1.0
            Parameter: x5 ScaleFactor: 10.0
            Parameter: z3 ScaleFactor: 1.0
            Parameter: x6 ScaleFactor: 10.0
            Parameter: z4 ScaleFactor: 1.0
            Parameter: x7 ScaleFactor: 10.0
            Parameter: z5 ScaleFactor: 1.0
         InputConstraints
            Parameter: x1 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x3 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x4 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x5 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x6 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x7 LowerBound: -100.0 UpperBound: 100.0
         PotentialObjectives:
            z6 z7 z8 x1 x3 z1 x4 z2 x5 z3
            x6 z4 x7 z5 x8 J1 x2
         Objectives
            Parameter: J1 Direction: minimize Weight: 1.0 ScaleFactor: 1e-3
         OutputConstraints
            Parameter: x8 LowerBound: -100.0 Weight: 1.0 ScaleFactor: 1.0 UpperBound: 100.0 Weight: 1.0 ScaleFactor: 1.0
            Parameter: x2 LowerBound: -100.0 Weight: 1.0 ScaleFactor: 1.0 UpperBound: 100.0 Weight: 1.0 ScaleFactor: 1.0

         OptimizePlan NewPlan
            DefaultUpperBound: 1.0E15
            UseScaling: yes
            OptimizeStep Step1
                  Technique: "Generalized Reduced Gradient - LSGRG2"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
                     MaxIterations: 100
            OptimizeStep Step2
                  Technique: "Hooke-Jeeves Direct Search Method"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
            Control: [
                  Step1
                  Step2
            ]

         # PLAN TO BE CONFIGURED BY ADVISOR:
         OptimizePlan PriorityRankedPlan
            Control: [
            ]
      End Optimization output28

      TaskPlan output28
         StopTaskPlanOnError: no
         Control: [
            Tcl
                  api_UnsetBestRunInfo [api_GetTaskName]
                  api_SetRunCounter [api_GetTaskName] 1
            End Tcl
            NewPlan
         ]
      End TaskPlan output28

      DataStorage output28
         Restore: no
         DataLog: "output28.db" Mode: append
         DataLookUp: "output28.db"
         MatchMode: Exact
         Levels: betteronly
         StoreGradRuns: yes
         StoreApproxRuns: yes
      End DataStorage output28

End Task output28

Task output35

      TaskHeader output35
         Version: 1.0
         Evaluation: taskplan
         ControlMode: user
         RunCounter: 1
         BoundsPolicy: adjustvalue
         CheckPoint: unknown
      End TaskHeader output35

      Inputs output35
         Parameter: x8 Type: real InitialValue: 2.0
         Parameter: x4 Type: real InitialValue: 4.0
         Parameter: x1 Type: real InitialValue: -10.0
         Parameter: x6 Type: real InitialValue: 2.0
         Parameter: x2 Type: real InitialValue: -10.0
         Parameter: x7 Type: real InitialValue: -2.0
         Parameter: z6 Type: real InitialValue: 1.8705959911958
         Parameter: z7 Type: real InitialValue: -1.58718411550964
         Parameter: z8 Type: real InitialValue: -1.44375507956317
         Parameter: z1 Type: real InitialValue: 8.44674964678718
         Parameter: z2 Type: real InitialValue: -9.05310287720698
         Parameter: z3 Type: real InitialValue: -0.484397005726966
         Parameter: z4 Type: real InitialValue: 3.06553246855229
         Parameter: z5 Type: real InitialValue: 1.64978849482124
      End Inputs output35

      Outputs output35
         Parameter: x5 Type: real
         Parameter: x3 Type: real
         Parameter: J2 Type: real
      End Outputs output35

      Calculations output35
         Calculation Calculation4
            Parameters
                  z6 x8 z7 z8 x1 x2 z1 x4 z2 z3
                  x6 z4 z5 x7 J2 x5 x3
            Statements
                  x3 = 2-x4
                  x5=(-4-x1*x7-x2*x8-x4*x6)/x3

                  J2=(x1-z1)^2+(x2-z2)^2+(x3-z3)^2+(x4-z4)^2+(x5-z5)^2+(x6-z6)^2+(x7-z7)^2+(x8-z8)^2
            End Statements
         End Calculation Calculation4
      End Calculations output35

      TaskProcess output35
         Control: [
            Calculation4
         ]
      End TaskProcess output35

      Optimization output35
         PotentialVariables:
            x8 x4 x1 x6 x2 x7 z6 z7 z8 z1
            z2 z3 z4 z5
         Variables:
            x8 x4 x1 x6 x2 x7
         VariableScaling
            Parameter: x8 ScaleFactor: 10.0
            Parameter: x4 ScaleFactor: 10.0
            Parameter: x1 ScaleFactor: 10.0
            Parameter: x6 ScaleFactor: 10.0
            Parameter: x2 ScaleFactor: 10.0
            Parameter: x7 ScaleFactor: 10.0
            Parameter: z6 ScaleFactor: 1.0
            Parameter: z7 ScaleFactor: 1.0
            Parameter: z8 ScaleFactor: 1.0
            Parameter: z1 ScaleFactor: 1.0
            Parameter: z2 ScaleFactor: 1.0
            Parameter: z3 ScaleFactor: 1.0
            Parameter: z4 ScaleFactor: 1.0
            Parameter: z5 ScaleFactor: 1.0
         InputConstraints
            Parameter: x8 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x4 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x1 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x6 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x2 LowerBound: -100.0 UpperBound: 100.0
            Parameter: x7 LowerBound: -100.0 UpperBound: 100.0
         PotentialObjectives:
            x8 x4 x1 x6 x2 x7 z6 z7 z8 z1
            z2 z3 z4 z5 x5 x3 J2
         Objectives
            Parameter: J2 Direction: minimize Weight: 1.0 ScaleFactor: 1e-3
         OutputConstraints
            Parameter: x5 LowerBound: -100.0 Weight: 1.0 ScaleFactor: 1.0 UpperBound: 100.0 Weight: 1.0 ScaleFactor: 1.0
            Parameter: x3 LowerBound: -100.0 Weight: 1.0 ScaleFactor: 1.0 UpperBound: 100.0 Weight: 1.0 ScaleFactor: 1.0

         OptimizePlan NewPlan
            DefaultUpperBound: 1.0E15
            UseScaling: yes
            OptimizeStep Step1
                  Technique: "Generalized Reduced Gradient - LSGRG2"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
                     MaxIterations: 100
            OptimizeStep Step2
                  Technique: "Hooke-Jeeves Direct Search Method"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
            Control: [
                  Step1
                  Step2
            ]

         # PLAN TO BE CONFIGURED BY ADVISOR:
         OptimizePlan PriorityRankedPlan
            Control: [
            ]
      End Optimization output35

      TaskPlan output35
         StopTaskPlanOnError: no
         Control: [
            Tcl
                  api_UnsetBestRunInfo [api_GetTaskName]
                  api_SetRunCounter [api_GetTaskName] 1
            End Tcl
            NewPlan
         ]
      End TaskPlan output35

      DataStorage output35
         Restore: no
         DataLog: "output35.db" Mode: append
         DataLookUp: "output35.db"
         MatchMode: Exact
         Levels: betteronly
         StoreGradRuns: yes
         StoreApproxRuns: yes
      End DataStorage output35

End Task output35

TaskProcess HeartDiopleCO
      Control: [
         Sequential [
            Parallel [ output28 output35 ]
            Calculation5
         ]
      ]

      SubTask output28
         InputToSubtask
            Send:
                  z1 z2 z3 z4 z5 z6 z7 z8
         OutputFromSubtask
            Receive:
                  J1
      End SubTask output28

      SubTask output35
         InputToSubtask
            Send:
                  z1 z2 z3 z4 z5 z6 z7 z8
         OutputFromSubtask
            Receive:
                  J2
      End SubTask output35

End TaskProcess HeartDiopleCO

Optimization HeartDiopleCO
      PotentialVariables:
         z6 z7 z8 z1 z2 z3 z4 z5
      Variables:
         z6 z7 z8 z1 z2 z3 z4 z5
      VariableScaling
         Parameter: z6 ScaleFactor: 10.0
         Parameter: z7 ScaleFactor: 10.0
         Parameter: z8 ScaleFactor: 10.0
         Parameter: z1 ScaleFactor: 10.0
         Parameter: z2 ScaleFactor: 10.0
         Parameter: z3 ScaleFactor: 10.0
         Parameter: z4 ScaleFactor: 10.0
         Parameter: z5 ScaleFactor: 10.0
      InputConstraints
         Parameter: z6 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z7 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z8 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z1 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z2 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z3 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z4 LowerBound: -100.0 UpperBound: 100.0
         Parameter: z5 LowerBound: -100.0 UpperBound: 100.0
      PotentialObjectives:
         z6 z7 z8 z1 z2 z3 z4 z5 J2 J1
         f6 f7 minF f8 f5
      Objectives
         Parameter: minF Direction: minimize Weight: 1.0 ScaleFactor: 1.0
      OutputConstraints
         Parameter: J2 UpperBound: 0.01 Weight: 1.0 ScaleFactor: 1.0
         Parameter: J1 UpperBound: 0.01 Weight: 1.0 ScaleFactor: 1.0
         Parameter: f6 LowerBound: 0.0 Weight: 1.0 ScaleFactor: 1.0
         Parameter: f7 LowerBound: 0.0 Weight: 1.0 ScaleFactor: 1.0
         Parameter: f8 LowerBound: 0.0 Weight: 1.0 ScaleFactor: 1.0
         Parameter: f5 LowerBound: 0.0 Weight: 1.0 ScaleFactor: 1.0

      OptimizePlan NewPlan
         DefaultUpperBound: 1.0E15
         UseScaling: yes
         OptimizeStep Step1
            Technique: "Sequential Quadratic Programming - DONLP"
            Prologue
                  RestoreBestSolution: no
                  RerunTask: no
            Epilogue
                  RestoreBestSolution: yes
                  RerunTask: no
            Options
                  Maxiter: 100
         Control: [
            Step1
         ]

      # PLAN TO BE CONFIGURED BY ADVISOR:
      OptimizePlan PriorityRankedPlan
         Control: [
         ]
End Optimization HeartDiopleCO

TaskPlan HeartDiopleCO
      StopTaskPlanOnError: no
      Control: [
         Tcl
            api_UnsetBestRunInfo [api_GetTaskName]
            api_SetRunCounter [api_GetTaskName] 1
         End Tcl
         NewPlan
      ]
End TaskPlan HeartDiopleCO

DataStorage HeartDiopleCO
      Restore: no
      DataLog: "HeartDiopleCO.db" Mode: append
      DataLookUp: "HeartDiopleCO.db"
      MatchMode: Exact
      Levels: all
      StoreGradRuns: yes
      StoreApproxRuns: yes
End DataStorage HeartDiopleCO

End Task HeartDiopleCO

另海沟的例子（较为简单）
MDOLVersion: 8.0
CompilerOptions: warn

Task RosenbrockvalleyCO

TaskHeader RosenbrockvalleyCO
      Version: 1.0
      Evaluation: taskplan
      ControlMode: user
      RunCounter: 1
      BoundsPolicy: adjustvalue
      CheckPoint: unknown
End TaskHeader RosenbrockvalleyCO

Inputs RosenbrockvalleyCO
      Parameter: z2 Type: real InitialValue: 0.0
      Parameter: z1 Type: real InitialValue: 0.0
End Inputs RosenbrockvalleyCO

Outputs RosenbrockvalleyCO
      Parameter: J1 Type: real
      Parameter: J2 Type: real
      Parameter: F Type: real
End Outputs RosenbrockvalleyCO

Task Outputy21

      TaskHeader Outputy21
         Version: 1.0
         Evaluation: taskplan
         ControlMode: user
         RunCounter: 1
         BoundsPolicy: adjustvalue
         CheckPoint: unknown
      End TaskHeader Outputy21

      Inputs Outputy21
         Parameter: z2 Type: real InitialValue: 0.0
         Parameter: x1 Type: real InitialValue: 0.0
         Parameter: x2 Type: real InitialValue: 0.0
         Parameter: z1 Type: real InitialValue: 0.0
      End Inputs Outputy21

      Outputs Outputy21
         Parameter: y21 Type: real
         Parameter: J1 Type: real
      End Outputs Outputy21

      Calculations Outputy21
         Calculation Calculation3
            Parameters
                  z2 x1 x2 z1 y21 J1
            Statements
                  y21=100*(x2-x1^2)^2

                  J1=(x1-z1)^2+(y21-z2)^2
            End Statements
         End Calculation Calculation3
      End Calculations Outputy21

      TaskProcess Outputy21
         Control: [
            Calculation3
         ]
      End TaskProcess Outputy21

      Optimization Outputy21
         PotentialVariables:
            z2 x1 x2 z1
         Variables:
            x1 x2
         VariableScaling
            Parameter: z2 ScaleFactor: 1.0
            Parameter: x1 ScaleFactor: 1.0
            Parameter: x2 ScaleFactor: 1.0
            Parameter: z1 ScaleFactor: 1.0
         InputConstraints
            Parameter: x1 LowerBound: -10.0 UpperBound: 10.0
            Parameter: x2 LowerBound: -10.0 UpperBound: 10.0
         PotentialObjectives:
            z2 x1 x2 z1 y21 J1
         Objectives
            Parameter: J1 Direction: minimize Weight: 1.0 ScaleFactor: 1.0

         OptimizePlan NewPlan
            DefaultUpperBound: 1.0E15
            UseScaling: yes
            OptimizeStep Step1
                  Technique: "Generalized Reduced Gradient - LSGRG2"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
                     MaxIterations: 100
                     ConvergenceEpsilon: 1e-6
            Control: [
                  Step1
            ]

         # PLAN TO BE CONFIGURED BY ADVISOR:
         OptimizePlan PriorityRankedPlan
            Control: [
            ]
      End Optimization Outputy21

      TaskPlan Outputy21
         StopTaskPlanOnError: no
         Control: [
            Tcl
                  api_SetRunCounter [api_GetTaskName] 1

                  api_UnsetBestRunInfo [api_GetTaskName]
            End Tcl
            NewPlan
         ]
      End TaskPlan Outputy21

      DataStorage Outputy21
         Restore: no
         DataLog: "Outputy21.db" Mode: append
         DataLookUp: "Outputy21.db"
         MatchMode: Exact
         Levels: betteronly
         StoreGradRuns: yes
         StoreApproxRuns: yes
      End DataStorage Outputy21

End Task Outputy21

Task OutputF

      TaskHeader OutputF
         Version: 1.0
         Evaluation: taskplan
         ControlMode: user
         RunCounter: 1
         BoundsPolicy: adjustvalue
         CheckPoint: unknown
      End TaskHeader OutputF

      Inputs OutputF
         Parameter: z2 Type: real InitialValue: 0.0
         Parameter: y21 Type: real InitialValue: 0.0
         Parameter: x1 Type: real InitialValue: 0.0
         Parameter: z1 Type: real InitialValue: 0.0
      End Inputs OutputF

      Outputs OutputF
         Parameter: J2 Type: real
         Parameter: F Type: real
      End Outputs OutputF

      Calculations OutputF
         Calculation Calculation4
            Parameters
                  z2 y21 x1 z1 J2 F
            Statements
                  F=y21+(1-x1)^2

                  J2=(x1-z1)^2+(y21-z2)^2
            End Statements
         End Calculation Calculation4
      End Calculations OutputF

      TaskProcess OutputF
         Control: [
            Calculation4
         ]
      End TaskProcess OutputF

      Optimization OutputF
         PotentialVariables:
            z2 y21 x1 z1
         Variables:
            y21 x1
         VariableScaling
            Parameter: z2 ScaleFactor: 1.0
            Parameter: y21 ScaleFactor: 1.0
            Parameter: x1 ScaleFactor: 1.0
            Parameter: z1 ScaleFactor: 1.0
         InputConstraints
            Parameter: y21 LowerBound: 0.0
            Parameter: x1 LowerBound: -10.0 UpperBound: 10.0
         PotentialObjectives:
            z2 y21 x1 z1 J2 F
         Objectives
            Parameter: J2 Direction: minimize Weight: 1.0 ScaleFactor: 1.0

         OptimizePlan NewPlan
            DefaultUpperBound: 1.0E15
            UseScaling: yes
            OptimizeStep Step1
                  Technique: "Generalized Reduced Gradient - LSGRG2"
                  Prologue
                     RestoreBestSolution: no
                     RerunTask: no
                  Epilogue
                     RestoreBestSolution: yes
                     RerunTask: no
                  Options
                     MaxIterations: 100
                     ConvergenceEpsilon: 1e-6
            Control: [
                  Step1
            ]

         # PLAN TO BE CONFIGURED BY ADVISOR:
         OptimizePlan PriorityRankedPlan
            Control: [
            ]
      End Optimization OutputF

      TaskPlan OutputF
         StopTaskPlanOnError: no
         Control: [
            NewPlan
            Tcl
                  api_UnsetBestRunInfo [api_GetTaskName]
                  api_SetRunCounter [api_GetTaskName] 1
            End Tcl
         ]
      End TaskPlan OutputF

      DataStorage OutputF
         Restore: no
         DataLog: "OutputF.db" Mode: append
         DataLookUp: "OutputF.db"
         MatchMode: Exact
         Levels: betteronly
         StoreGradRuns: yes
         StoreApproxRuns: yes
      End DataStorage OutputF

End Task OutputF

TaskProcess RosenbrockvalleyCO
      Control: [
         Parallel [ Outputy21 OutputF ]
      ]

      SubTask Outputy21
         InputToSubtask
            Send:
                  z1 z2
         OutputFromSubtask
            Receive:
                  J1
      End SubTask Outputy21

      SubTask OutputF
         InputToSubtask
            Send:
                  z1 z2
         OutputFromSubtask
            Receive:
                  F J2
      End SubTask OutputF

End TaskProcess RosenbrockvalleyCO

Optimization RosenbrockvalleyCO
      PotentialVariables:
         z2 z1
      Variables:
         z2 z1
      VariableScaling
         Parameter: z2 ScaleFactor: 1.0
         Parameter: z1 ScaleFactor: 1.0
      InputConstraints
         Parameter: z2 LowerBound: 0.0
         Parameter: z1 LowerBound: -10.0 UpperBound: 10.0
      PotentialObjectives:
         z2 z1 J1 J2 F
      Objectives
         Parameter: F Direction: minimize Weight: 1.0 ScaleFactor: 1.0
      OutputConstraints
         Parameter: J1 UpperBound: 0.01 Weight: 1.0 ScaleFactor: 1.0
         Parameter: J2 UpperBound: 0.01 Weight: 1.0 ScaleFactor: 1.0

      OptimizePlan NewPlan
         DefaultUpperBound: 1.0E15
         UseScaling: yes
         OptimizeStep Step1
            Technique: "Sequential Quadratic Programming - DONLP"
            Prologue
                  RestoreBestSolution: no
                  RerunTask: no
            Epilogue
                  RestoreBestSolution: yes
                  RerunTask: no
            Options
                  Maxiter: 100
         Control: [
            Step1
         ]

      # PLAN TO BE CONFIGURED BY ADVISOR:
      OptimizePlan PriorityRankedPlan
         Control: [
         ]
End Optimization RosenbrockvalleyCO

TaskPlan RosenbrockvalleyCO
      StopTaskPlanOnError: no
      Control: [
         Tcl
            api_UnsetBestRunInfo [api_GetTaskName]
            api_SetRunCounter [api_GetTaskName] 1
         End Tcl
         NewPlan
      ]
End TaskPlan RosenbrockvalleyCO

DataStorage RosenbrockvalleyCO
      Restore: no
      DataLog: "RosenbrockvalleyCO.db" Mode: append
      DataLookUp: "RosenbrockvalleyCO.db"
      MatchMode: Exact
      Levels: all
      StoreGradRuns: yes
      StoreApproxRuns: yes
End DataStorage RosenbrockvalleyCO

End Task RosenbrockvalleyCO
//
结果为：
End time: Mon May 29 19:22:44 WST 2006

Task: RosenbrockvalleyCO
  Total runs:    7
  Feasible runs: 7
  Infeasible runs: 0
  Failed runs:    0
  Database file: RosenbrockvalleyCO.db

  Optimization Plan: NewPlan
Executed between RunCounter 1 and 7 (7 runs)

Techniques used:
      Step1: Sequential Quadratic Programming - DONLP

Best design:             currently             previously
--------------------------------------------------------------------
      RunCounter          5                   1
      ObjectiveAndPenalty 2.49999998050393e-009  1.0
      Objective          2.49999998050393e-009  1.0
      Penalty             0.0                   0.0

Best design ObjectiveAndPenalty value improved by
0.9999999975 (100.0%) after executing this Optimization Plan

Best design parameter values:

      z2 = 1.00000000002676e-006
      z1 = 0.999950000000194
      J1 = 6.25766180395944e-008
      J2 = 1.00000000005352e-012
      F  = 2.49999998050393e-009
第二组为；
Execution Results Summary

Start time: Mon May 29 19:25:55 WST 2006
End time: Mon May 29 19:26:29 WST 2006

Task: RosenbrockvalleyCO
  Total runs:    7
  Feasible runs: 7
  Infeasible runs: 0
  Failed runs:    0
  Database file: RosenbrockvalleyCO.db

  Optimization Plan: NewPlan
Executed between RunCounter 1 and 7 (7 runs)

Techniques used:
      Step1: Sequential Quadratic Programming - DONLP

Best design:             currently             previously
--------------------------------------------------------------------
      RunCounter          5                   1
      ObjectiveAndPenalty 2.49999998050393e-009  1.0
      Objective          2.49999998050393e-009  1.0
      Penalty             0.0                   0.0

Best design ObjectiveAndPenalty value improved by
0.9999999975 (100.0%) after executing this Optimization Plan

Best design parameter values:

      z2 = 1.00000000002676e-006
      z1 = 0.999950000000194
      J1 = 1.07944409357371e-007
      J2 = 1.00000000005352e-012
      F  = 2.49999998050393e-009

strongsailor · 发表于 2006-5-30 22:40:02

赞一个,楼主朋友费心了!

strongsailor · 发表于 2006-5-31 09:03:03

CO优化的初衷就是减少计算量,但是通过上面的HeartDIople和其它我所做的其它CO例子来看,如何在isight中进行设置,以提高收敛效率.望大家能够多多提出意见.

rabitzh · 发表于 2006-5-31 22:27:01

我看HeartDiople问题的结果数据就是感觉到其收敛的速度很慢，从系统迭代次数来看也是这样的，可能是SQP——ＤＯＮＬＰ高级参数设置方面有问题，我用的是默认的设置，各位同仁又研究过其参数影响的没有？
很急待高级参数设置研究结果。

显示全部楼层 · 发表于 2006-11-20 08:27:00

提示: 作者被禁止或删除内容自动屏蔽

账号		自动登录	找回密码
密码			注册

账号		自动登录	找回密码
密码			立即注册

fenfen8686 该用户已被删除	发表于 2006-11-20 08:27:00 \| 显示全部楼层来自北京理工大学提示: 作者被禁止或删除内容自动屏蔽
fenfen8686 该用户已被删除
	回复赞不支持使用道具举报显身卡

[iSIGHT] 分享我的协同优化的成果——HeartDiople和海沟问题

相关帖子

麻烦各同仁都研究研究HeartDiople