search

mlr-org / mlrMBO

Toolbox for Bayesian Optimization and Model-Based Optimization in R
https://mlrmbo.mlr-org.com
Other
187 stars 47 forks source link

Error: Error in as.data.frame.OptPathDF(opt.path, include.rest = FALSE) : No elements where selected (via 'dob' and 'eol')! #516

Open swaheera opened 3 years ago

swaheera commented 3 years ago

I recently got this error while trying to perform multiobjective constrained optimization with the mlrMBO library:


library(mlrMBO)

obj.fn = makeMultiObjectiveFunction(
    name = "My test function",
    fn = function(x1, x2, x3, x4) {
        var_1 <- sin(x1 + x2)
        var_2 <- cos(x1 - x2)
        var_3 <- x1 + x4
        var_4 <- x3 + x4 -7
        goal_1 = sum(var_1 + var_2 + var_3 + var_4)
        goal_2 = var_1 + var_2 - var_3 + var_4
        goal_3 = var_1 + var_2 - var_3 + 2*var_4

        return(c(goal_1, goal_2, goal_3))

    },
    n.objectives = 3L,
    #define acceptable ranges
    par.set = makeParamSet(
        makeNumericParam("x1", lower = 20, upper = 40),
        makeNumericParam("x2", lower = 30, upper = 45),
        makeNumericParam("x3", lower = 10, upper = 20),
        makeNumericParam("x4", lower = 10, upper = 50)
        #define constraints
        , forbidden = expression(x2 >x1 | x3 > x4)
    ),

    minimize=rep(TRUE,3)
)

#create control gird
 control=makeMBOControl(propose.points=1, final.method="best.predicted", final.evals=10)
 control=setMBOControlTermination(control, iters=10)
 control=setMBOControlInfill(control, crit=makeMBOInfillCritEI())

#perform optimization
lrn=makeMBOLearner(control, obj.fun)

res = mbo(obj.fun, design = NULL, learner = lrn, control = ctrl, show.info = TRUE)

#error
Warning in generateDesign(n.params * 4L, par.set, fun = lhs::maximinLHS) :
  generateDesign could only produce 0 points instead of 16!
Computing y column(s) for design. Not provided.

Error in as.data.frame.OptPathDF(opt.path, include.rest = FALSE) : 
  No elements where selected (via 'dob' and 'eol')!

I am a bit confused. Where exactly do I need to specify "dob" and "eol"?

Thanks

jakob-r commented 3 years ago

Just by looking at the constraints it looks like they are too narrow (and partly don't make sense) Forbidden: x2 >x1 | x3 > x4 In other words: x2 should always be smaller or equal then x1 AND x3 always has to be smaller then x4. Having this in mind it does not make sense to have a higher lower bound for x2 then for x1. Probably there is no feasible point found in the initial design.

swaheera commented 3 years ago

I tried removing the constraints all together, but now I get a different error:


library(mlrMBO)

obj.fn = makeMultiObjectiveFunction(
    name = "My test function",
    fn = function(x1, x2, x3, x4) {
        var_1 <- sin(x1 + x2)
        var_2 <- cos(x1 - x2)
        var_3 <- x1 + x4
        var_4 <- x3 + x4 -7
        goal_1 = sum(var_1 + var_2 + var_3 + var_4)
        goal_2 = var_1 + var_2 - var_3 + var_4
        goal_3 = var_1 + var_2 - var_3 + 2*var_4

        return(c(goal_1, goal_2, goal_3))

    },
    n.objectives = 3L,
    #define acceptable ranges
    par.set = makeParamSet(
        makeNumericParam("x1", lower = 20, upper = 40),
        makeNumericParam("x2", lower = 30, upper = 45),
        makeNumericParam("x3", lower = 10, upper = 20),
        makeNumericParam("x4", lower = 10, upper = 50)

    ),

    minimize=rep(TRUE,3)
)

#create control gird
 control=makeMBOControl(propose.points=1, final.method="best.predicted", final.evals=10)
 control=setMBOControlTermination(control, iters=10)
 control=setMBOControlInfill(control, crit=makeMBOInfillCritEI())

#perform optimization
lrn=makeMBOLearner(control, obj.fun)

res = mbo(obj.fun, design = NULL, learner = lrn, control = ctrl, show.info = TRUE)

Error:

Error in checkClass(x, classes, ordered, null.ok) : 
  object 'obj.fun' not found

Is there any way to fix this? Thanks

jakob-r commented 3 years ago

You named your function obj.fn and not obj.fun

swaheera commented 3 years ago

Thank you for the correction - I found another typo : "ctrl" vs "control". I fixed both of these errors but now I have a new error.

swaheera commented 3 years ago 
library(mlrMBO)

obj.fn = makeMultiObjectiveFunction(
    name = "My test function",
    fn = function(x1, x2, x3, x4) {
        var_1 <- sin(x1 + x2)
        var_2 <- cos(x1 - x2)
        var_3 <- x1 + x4
        var_4 <- x3 + x4 -7
        goal_1 = sum(var_1 + var_2 + var_3 + var_4)
        goal_2 = var_1 + var_2 - var_3 + var_4
        goal_3 = var_1 + var_2 - var_3 + 2*var_4

        return(c(goal_1, goal_2, goal_3))

    },
    n.objectives = 3L,
    #define acceptable ranges
    par.set = makeParamSet(
        makeNumericParam("x1", lower = 20, upper = 40),
        makeNumericParam("x2", lower = 30, upper = 45),
        makeNumericParam("x3", lower = 10, upper = 20),
        makeNumericParam("x4", lower = 10, upper = 50)
        #define constraints
        , forbidden = expression(x2 >x1 | x3 > x4)
    ),

    minimize=rep(TRUE,3)
)

#create control gird
control=makeMBOControl(propose.points=1, final.method="best.predicted",  final.evals=10)
control=setMBOControlTermination(control, iters=10)
control=setMBOControlInfill(control, crit=makeMBOInfillCritEI())

#perform optimization
lrn=makeMBOLearner(control, obj.fn)

res = mbo(obj.fn, design = NULL, learner = lrn, control = control, show.info = TRUE)

I got the following error:

Warning in generateDesign(n.params * 4L, par.set, fun = lhs::maximinLHS) :
  generateDesign could only produce 15 points instead of 16!
Error in checkStuff(fun, design, learner, control) : 
  Objective function has 3 objectives, but the control object assumes 1.

I tried to fix this by changing the number of objectives:

#create control gird
control=makeMBOControl(propose.points=1, final.method="best.predicted",  n.objectives = 3L, final.evals=10)
control=setMBOControlTermination(control, iters=10)
control=setMBOControlInfill(control, crit=makeMBOInfillCritEI())

#perform optimization
lrn=makeMBOLearner(control, obj.fn)

res = mbo(obj.fn, design = NULL, learner = lrn, control = control, show.info = TRUE)

But then I got a new error: 

Error in checkStuff(fun, design, learner, control) : 
  Setting of final.method and final.evals for multi-objective optimization not supported at the moment.

Do you have any idea why this error is being produced?

Thank you so much for your help!

Thanks

jakob-r commented 3 years ago

In MOO there is no final best but only a set of points that are pareto optimal.

You have to change makeMBOControl

control=makeMBOControl(propose.points=1)
swaheera commented 3 years ago

Hello Dr. Richter,

Thank you for your reply.

I removed the "best.predicted" statement, but the code is still not working:

library(mlrMBO)

obj.fn = makeMultiObjectiveFunction(
    name = "My test function",
    fn = function(x1, x2, x3, x4) {
        var_1 <- sin(x1 + x2)
        var_2 <- cos(x1 - x2)
        var_3 <- x1 + x4
        var_4 <- x3 + x4 -7
        goal_1 = sum(var_1 + var_2 + var_3 + var_4)
        goal_2 = var_1 + var_2 - var_3 + var_4
        goal_3 = var_1 + var_2 - var_3 + 2*var_4

        return(c(goal_1, goal_2, goal_3))

    },
    n.objectives = 3L,
    #define acceptable ranges
    par.set = makeParamSet(
        makeNumericParam("x1", lower = 20, upper = 40),
        makeNumericParam("x2", lower = 30, upper = 45),
        makeNumericParam("x3", lower = 10, upper = 20),
        makeNumericParam("x4", lower = 10, upper = 50)
        #define constraints
        , forbidden = expression(x2 >x1 | x3 > x4)
    ),

    minimize=rep(TRUE,3)
)

#create control gird
control=makeMBOControl(propose.points=1,  n.objectives = 3L, final.evals=10)

#perform optimization
lrn=makeMBOLearner(control, obj.fn)

res = mbo(obj.fn, design = NULL, learner = lrn, control = control, show.info = TRUE)

But this produces the following error

Error in checkStuff(fun, design, learner, control) : 
  Setting of final.method and final.evals for multi-objective optimization not supported at the moment.

As a result of this error, I tried to remove "final.evals":


#create control gird
control=makeMBOControl(propose.points=1,  n.objectives = 3L)

#perform optimization
lrn=makeMBOLearner(control, obj.fn)

res = mbo(obj.fn, design = NULL, learner = lrn, control = control, show.info = TRUE)

But now I get a different error (even though I have specified there are 3 objectives):

Warning in generateDesign(n.params * 4L, par.set, fun = lhs::maximinLHS) :
  generateDesign could only produce 15 points instead of 16!
Error in checkStuff(fun, design, learner, control) : 
  Objective function has 3 objectives, but the control object assumes 1.

If you have time, can you please try running this code on your computer and see if you can get it to work? I have been trying to get this to work for a while, but without any results.

Your Help Is Greatly Appreciated, Thanks

swaheera commented 3 years ago

Hello Dr. Richter,,

Can you please take a look at this if you have some time?

Thanks