CODEWITHSUNDEEP
rrandom
jeffbraveheartgroup
jeffbraveheartgroup

Reputation: 1

Random number generation in R without a contiguous repeat

I want to test the useful life of two parts where one (part A) moves within the other (part B). There are 10 possible distinct "destinations" for Part A to land. Useful life is defined as 15,000 movements.

I've tried to generate a vector of "destinations" and eliminate having the same "destination" appear back-to-back. I can generate the vector but cannot figure out how to eliminate a back-to-back repeat. With acceptable output I can program a motor to step through each "destination" until useful life and test other mechanical attributes of both parts.

Here's my code:


    iterations <- (50) ##just to test the code

    LSVmin <- (1)
    LSVmax <-(10)

    movementV <- matrix(data=0, nrow=iterations, ncol=1) ##create vector for movement values 

    movementV[1,] <- 10

    i <- 2
    for (i in 2:iterations) {

        movement1 <- (floor(runif(1, min=LSVmin, max=LSVmax)))
        movementV[i, ] <- movement1
   
        if (identical(movementV[i, ], movementV[i-1, ]) > 0) 

                movementV[i, ] <- movement1
   
        else  {

                movement2 <- (floor(runif(1, min=LSVmin, max=LSVmax))) ##no provision for repeats 
                movementV[i, ] <- movement2
              }
    i<- (i + 1)  ### iterate the loop
     }

I've tried a number of methods to compare subsequent values but I don't know how to establish a variable within the if...else routine and not lose it when the loop iterates.

Upvotes: 0

Views: 87

Answers (2)

Maxime Garnault
Maxime Garnault

Reputation: 13

If I correctly understand what you want, I suggest you use the sample() function as follow:

iterations=50
LSVmin=1
LSVmax=10
posssibleChoices=c(LSVmin:LSVmax)

movement=rep(NA,iterations) 
movement[1]=10

for(i in 2:iterations){
  movement[i]=sample(posssibleChoices[-which(posssibleChoices==movement[i-1])],1)
}

movement
    [1] 10  8  4  5  9  1 10  9  1  8  3  6  9  8  6  4  8 10  6  7  5  7 10  8  4  6  8
    [28]  5  7  1  3  4  9  2  4  9  4  6  2  5  9  6  7  5 10  4  2  6  7  5

(1) Start with vector full of NA

(2) Initiate the first value of this vector to 10 (I kept this step because that is maybe something you want or need)

(3) For loop from 2 to iterations, sampling the new value in a vector containing all your possible choices (posssibleChoices) but not containing the lastest value from movement (movement[i-1])

PS: no need to increment i in the for loop, as it does it on its own (which is not the case for while loops of course).

Upvotes: 0

heds1
heds1

Reputation: 3438

Is this what you're after? If you want integers you can more easily use sample rather than floor(runif(...)). For each repetition, we sample a value from 1 to 10, then check that is not the same as the value in the previous index.

vals <- rep(0, 50)
for (i in seq_along(vals)) {
    new <- sample(10, 1, replace = TRUE)
    while (new == vals[max(i-1, 1)]) {
        new <- sample(10, 1, replace = TRUE)
    }
    vals[i] <- new
}

vals
#  [1]  5  1  2  6  3  4  6  7  5  4  8  9  8  4  2  8  1  4  8  2  4  9  4  7  2 10  3  2  9 10  3  8  6 10  1  5  3  1 10  1  5 10  3  7  1 10  2
# [48]  3  1  3

Upvotes: 1

Related Questions