how to force succesive variable to be of same value? - optimization

I have a binary variable y[k][t], where k = 1..3 (machines) and t = 1..10 (time). Variable Y is 1 if machine is active and 0 otherwise.
In the optimization if machine 1 is active in period 1 e.g. Y[1][1] = 1, i want the machine to continue to operate for at least 3 time periods. i.e. Y[1][1] = Y[1][2] = Y[1][3] = Y[1][4] = 1.
I only want the succesive variable for t+1,t+2,t+3 to be same as t if it is active.
how can i do that in cplex studio?

This is sometimes called a minimum run length. There are different approaches to handle this. Say we have x(t) as our binary variables indicating if the machine is operating at time t. The first thing is to introduce a binary variable start(t) that indicates when a run starts. This is what I mean:
t 1 2 3 4 5 6 7 8
x 0 1 1 1 0 1 1 0
start 0 1 0 0 0 1 0 0
A run starts when x(t-1)=0 and x(t)=1, or more formally:
start(t) = x(t)*(1-x(t-1))
This is nonlinear. We can linearize this using:
start(t) <= x(t)
start(t) <= 1-x(t-1)
start(t) >= x(t)-x(t-1)
Often we just use the bound:
start(t) >= x(t)-x(t-1)
Next we need:
start(t) = 1 ==> x(t)...x(t+K-1) = 1
where K is the minimum run length.
This can be modeled as:
x(t+1) >= start(t)
...
x(t+K-1) >= start(t)
(we already know that x(t)=1 if start(t)=1).

what Erwin wrote about minimum run length is fine. If you rely on logical constraints that are available in CPLEX the model is a bit easier:
range K=1..3;
range T=1..10;
dvar boolean Y[K][T];
dvar boolean start[K][T];
subject to
{
forall(k in K) start[k][1]==Y[k][1];
forall(k in K,t in T:t!=1) start[k][t] == ((Y[k][t]==1) && (Y[k][t-1]==0));
forall(k in K) forall(l in 1..3)
forall(t in T:(t+l) in T) (start[k][t]==1) => (Y[k][t+l]==1);
}
But what could lead to long solve time if you grow the time horizon is that we enumerate time. Within CPLEX and OPL you may also use CPOptimizer and its dedicated scheduling concepts : intervals.
Then you would write
using CP;
range K=1..3;
range T=1..10;
int nbMaxIntervals=4;
dvar interval itvs[K][1..nbMaxIntervals] optional in T size 3..10;
subject to
{
forall(k in K) forall(i in 1..nbMaxIntervals-1)
endBeforeStart(itvs[k][i],itvs[k][i+1]);
}
What makes sure that you are on for at least 3 time periods is
size 3..10;
NB: More about CPOptimizer

Related

How to convert the following if conditions to Linear integer programming constraints?

These are the conditions:
if(x > 0)
{
y >= a;
z <= b;
}
It is quite easy to convert the conditions into Linear Programming constraints if x were binary variable. But I am not finding a way to do this.
You can do this in 2 steps
Step 1: Introduce a binary dummy variable
Since x is continuous, we can introduce a binary 0/1 dummy variable. Let's call it x_positive
if x>0 then we want x_positive =1. We can achieve that via the following constraint, where M is a very large number.
x < x_positive * M
Note that this forces x_positive to become 1, if x is itself positive. If x is negative, x_positive can be anything. (We can force it to be zero by adding it to the objective function with a tiny penalty of the appropriate sign.)
Step 2: Use the dummy variable to implement the next 2 constraints
In English: if x_positive = 1, then y >= a
However, if x_positive = 0, y can be anything (y > -inf)
y > a - M (1 - x_positive)
Similarly,
if x_positive = 1, then z <= b
z <= b + M * (1 - x_positive)
Both the linear constraints above will kick in if x>0 and will be trivially satisfied if x <=0.

Running time of nested while loops

Function f(n)
s = 0
i = 1
while i < 7n^1/2 do
j = i
while j > 5 do
s = s + i -j
j = j -2
end
i = 5i
end
return s
end f
I am trying to solve the running time for big theta with the code above. I have been looking all over the place for something to help me with an example, but everything is for loops or only one while loop. How would you go about this problem with nested while loops?
Let's break this down into two key points:
i starts from 1, and is self-multiplied by 5, until it is greater than or equal to 7 sqrt(n). This is an exponential increase with logarithmic number of steps. Thus we can change the code to the following equivalent:
m = floor(log(5, 7n^(1/2)))
k = 0
while k < m do
j = 5^k
// ... inner loop ...
end
For each iteration of the outer loop, j starts from i, and decreases in steps of 2, until it is less than or equal to 5. Note that in the first execution of the outer loop i = 1, and in the second i = 5, so the inner loop is not executed until the third iteration. The loop limit means that the final value of j is 7 if k is odd, and 6 if even (you can check this with pen and paper).
Combining the above steps, we arrive at:
First loop will do 7 * sqrt(n) iterations. Exponent 1/2 is the same as sqrt() of a number.
Second loop will run m - 2 times since first two values of i are 1 and 5 respectively, not passing the comparison.
i is getting an increment of 5i.
Take an example where n = 16:
i = 1, n = 16;
while( i < 7 * 4; i *= 5 )
//Do something
First value of i = 1. It runs 1 time. Inside loop will run 0 times.
Second value of i = 5. It runs 2 times. Inside loop will run 0 times.
Third value of i = 25. It runs 3 times. Inside loop will run 10 times.
Fourth value of i = 125. It stops.
Outer iterations are n iterations while inner iterations are m iterations, which gives O( 7sqrt(n) * (m - 2) )
IMO, is complex.

Why is code like i=i*2 considered O(logN) when in a loop?

Why because of the i=i*2 is the runtime of the loop below considered O(logN)?
for (int i = 1; i <= N;) {
code with O(1);
i = i * 2;
}
Look at 1024 = 210. How many times do you have to double the number 1 to get 1024?
Times 1 2 3 4 5 6 7 8 9 10
Result 2 4 8 16 32 64 128 256 512 1024
So you would have to run your doubling loop ten times to get 210 And in general, you have to run your doubling loop n times to get 2n. But what is n? It's the log2 2n, so in general if n is some power of 2, the loop has to run log2n times to reach it.
To have the algorithm in O(logN), for any N it would need (around) log N steps. In the example of N=32 (where log 32 = 5) this can be seen:
i = 1 (start)
i = 2
i = 4
i = 8
i = 16
i = 32 (5 iterations)
i = 64 (abort after 6 iterations)
In general, after x iterations, i=2^x holds. To reach i>N you need x = log N + 1.
PS: When talking about complexities, the log base (2, 10, e, ...) is not relevant. Furthermore, it is not relevant if you have i <= N or i < N as this only changes the number of iterations by one.
You can prove it pretty simply.
Claim:
For the tth (0 base) iteration, i=2^t
Proof, by induction.
Base: 2^0 = 1, and indeed in the first iteration, i=1.
Step: For some t+1, the value of i is 2*i(t) (where i(t) is the value of i in the t iteration). From Induction Hypothesis we know that i(t)=2^t, and thus i(t+1) = 2*i(t) = 2*2^t = 2^(t+1), and the claim holds.
Now, let's examine our stop criteria. We iterate the loop while i <= N, and from the above claim, that means we iterate while 2^t <= N. By Doing log_2 on both sides, we get log_2(2^t) <= log_2(N), and since log_2(2^t) = t, we get that we iterate while t <= log_2(N) - so we iterate Theta(log_2(N)) times. (And that concludes the proof).
i starts in 1. In each iteration you multiply i by 2, so in the K-th iteration, i will be 2K-1.
After a number K of iterations, 2K-1 will be bigger than (or to) N.
this means N ≤ 2K-1
this means log2(N) ≤ K-1
K-1 will be the number of iterations your loop will run, and since K-1 is greater or equal to log(N), your algorithm is logarithmic.

GLPK - Minimize a variable value

I'm working in virtual network embedding, and I'm creating a model with glpk to embed the networks.
I have this following objective function:
minimize cost: sum{(i,j) in VEdges} sum{u in SNodes, v in SNodes} weight[u,v] * fw[i,j,u,v] * secSupEdge[u,v] + sum{u in SNodes, v in SNodes} r[u,v] * secSupEdge[u,v];
Then I have the following two restrictions (among others)
s.t. relConst2{(i,j) in VEdges, u in AllNodes, v in AllNodes}: bwDem[i,j] * phiw[i,j,u,v] >= fw[i,j,u,v];
s.t. linkSecConst0{(i,j) in VEdges, u in SNodes, v in SNodes}: phiw[i,j,u,v] * secDemEdge[i,j] <= secSupEdge[u,v];
"phiw" is a binary variable
"fw" and "r" are variables that take any value >= 0
all the others ("weight", "bwDem", "secDemEdges", "secSupEdge") are just params
I want to relate phiw with fw. When fw > 0, phiw should take the value 1. When fw == 0, phiw should take the value 0.
Normally it does what I want, but sometimes phiw takes the value 1 when fw has the value 0, which is not what I want. This happens because the restrictions are met:
Example 1:
s.t. relConst2: 4 * 1 >= 0
s.t. linkSecConst0: 1 * 2 <= 2
Is there a way to minimize the value of phiw variable but not putting it in the objective function? Or putting it in the objective function but not changing the value of the result neither the value of all other variables?
The question is about minimizing phiw, however, the description of the problem suggests that what you want to do is link the values of phiw and fw, and specifically to have phiw = 1 when fw > 0, and phiw = 0 otherwise (i.e., fw = 0).
I would suggest that you add constraint that directly maps the conditional on fw to the value of phiw, such as:
s.t. LinkConstraint { (i,j) in VEdges, u in AllNodes, v in AllNodes }:
if fw[i,j,u,v] > 0 then 1 else 0 = phiw[i,j,u,v] ;

Recognizing when to use the modulus operator

I know the modulus (%) operator calculates the remainder of a division. How can I identify a situation where I would need to use the modulus operator?
I know I can use the modulus operator to see whether a number is even or odd and prime or composite, but that's about it. I don't often think in terms of remainders. I'm sure the modulus operator is useful, and I would like to learn to take advantage of it.
I just have problems identifying where the modulus operator is applicable. In various programming situations, it is difficult for me to see a problem and realize "Hey! The remainder of division would work here!".
Imagine that you have an elapsed time in seconds and you want to convert this to hours, minutes, and seconds:
h = s / 3600;
m = (s / 60) % 60;
s = s % 60;
0 % 3 = 0;
1 % 3 = 1;
2 % 3 = 2;
3 % 3 = 0;
Did you see what it did? At the last step it went back to zero. This could be used in situations like:
To check if N is divisible by M (for example, odd or even)
or
N is a multiple of M.
To put a cap of a particular value. In this case 3.
To get the last M digits of a number -> N % (10^M).
I use it for progress bars and the like that mark progress through a big loop. The progress is only reported every nth time through the loop, or when count%n == 0.
I've used it when restricting a number to a certain multiple:
temp = x - (x % 10); //Restrict x to being a multiple of 10
Wrapping values (like a clock).
Provide finite fields to symmetric key algorithms.
Bitwise operations.
And so on.
One use case I saw recently was when you need to reverse a number. So that 123456 becomes 654321 for example.
int number = 123456;
int reversed = 0;
while ( number > 0 ) {
# The modulus here retrieves the last digit in the specified number
# In the first iteration of this loop it's going to be 6, then 5, ...
# We are multiplying reversed by 10 first, to move the number one decimal place to the left.
# For example, if we are at the second iteration of this loop,
# reversed gonna be 6, so 6 * 10 + 12345 % 10 => 60 + 5
reversed = reversed * 10 + number % 10;
number = number / 10;
}
Example. You have message of X bytes, but in your protocol maximum size is Y and Y < X. Try to write small app that splits message into packets and you will run into mod :)
There are many instances where it is useful.
If you need to restrict a number to be within a certain range you can use mod. For example, to generate a random number between 0 and 99 you might say:
num = MyRandFunction() % 100;
Any time you have division and want to express the remainder other than in decimal, the mod operator is appropriate. Things that come to mind are generally when you want to do something human-readable with the remainder. Listing how many items you could put into buckets and saying "5 left over" is good.
Also, if you're ever in a situation where you may be accruing rounding errors, modulo division is good. If you're dividing by 3 quite often, for example, you don't want to be passing .33333 around as the remainder. Passing the remainder and divisor (i.e. the fraction) is appropriate.
As #jweyrich says, wrapping values. I've found mod very handy when I have a finite list and I want to iterate over it in a loop - like a fixed list of colors for some UI elements, like chart series, where I want all the series to be different, to the extent possible, but when I've run out of colors, just to start over at the beginning. This can also be used with, say, patterns, so that the second time red comes around, it's dashed; the third time, dotted, etc. - but mod is just used to get red, green, blue, red, green, blue, forever.
Calculation of prime numbers
The modulo can be useful to convert and split total minutes to "hours and minutes":
hours = minutes / 60
minutes_left = minutes % 60
In the hours bit we need to strip the decimal portion and that will depend on the language you are using.
We can then rearrange the output accordingly.
Converting linear data structure to matrix structure:
where a is index of linear data, and b is number of items per row:
row = a/b
column = a mod b
Note above is simplified logic: a must be offset -1 before dividing & the result must be normalized +1.
Example: (3 rows of 4)
1 2 3 4
5 6 7 8
9 10 11 12
(7 - 1)/4 + 1 = 2
7 is in row 2
(7 - 1) mod 4 + 1 = 3
7 is in column 3
Another common use of modulus: hashing a number by place. Suppose you wanted to store year & month in a six digit number 195810. month = 195810 mod 100 all digits 3rd from right are divisible by 100 so the remainder is the 2 rightmost digits in this case the month is 10. To extract the year 195810 / 100 yields 1958.
Modulus is also very useful if for some crazy reason you need to do integer division and get a decimal out, and you can't convert the integer into a number that supports decimal division, or if you need to return a fraction instead of a decimal.
I'll be using % as the modulus operator
For example
2/4 = 0
where doing this
2/4 = 0 and 2 % 4 = 2
So you can be really crazy and let's say that you want to allow the user to input a numerator and a divisor, and then show them the result as a whole number, and then a fractional number.
whole Number = numerator/divisor
fractionNumerator = numerator % divisor
fractionDenominator = divisor
Another case where modulus division is useful is if you are increasing or decreasing a number and you want to contain the number to a certain range of number, but when you get to the top or bottom you don't want to just stop. You want to loop up to the bottom or top of the list respectively.
Imagine a function where you are looping through an array.
Function increase Or Decrease(variable As Integer) As Void
n = (n + variable) % (listString.maxIndex + 1)
Print listString[n]
End Function
The reason that it is n = (n + variable) % (listString.maxIndex + 1) is to allow for the max index to be accounted.
Those are just a few of the things that I have had to use modulus for in my programming of not just desktop applications, but in robotics and simulation environments.
Computing the greatest common divisor
Determining if a number is a palindrome
Determining if a number consists of only ...
Determining how many ... a number consists of...
My favorite use is for iteration.
Say you have a counter you are incrementing and want to then grab from a known list a corresponding items, but you only have n items to choose from and you want to repeat a cycle.
var indexFromB = (counter-1)%n+1;
Results (counter=indexFromB) given n=3:
`1=1`
`2=2`
`3=3`
`4=1`
`5=2`
`6=3`
...
Best use of modulus operator I have seen so for is to check if the Array we have is a rotated version of original array.
A = [1,2,3,4,5,6]
B = [5,6,1,2,3,4]
Now how to check if B is rotated version of A ?
Step 1: If A's length is not same as B's length then for sure its not a rotated version.
Step 2: Check the index of first element of A in B. Here first element of A is 1. And its index in B is 2(assuming your programming language has zero based index).
lets store that index in variable "Key"
Step 3: Now how to check that if B is rotated version of A how ??
This is where modulus function rocks :
for (int i = 0; i< A.length; i++)
{
// here modulus function would check the proper order. Key here is 2 which we recieved from Step 2
int j = [Key+i]%A.length;
if (A[i] != B[j])
{
return false;
}
}
return true;
It's an easy way to tell if a number is even or odd. Just do # mod 2, if it is 0 it is even, 1 it is odd.
Often, in a loop, you want to do something every k'th iteration, where k is 0 < k < n, assuming 0 is the start index and n is the length of the loop.
So, you'd do something like:
int k = 5;
int n = 50;
for(int i = 0;i < n;++i)
{
if(i % k == 0) // true at 0, 5, 10, 15..
{
// do something
}
}
Or, you want to keep something whitin a certain bound. Remember, when you take an arbitrary number mod something, it must produce a value between 0 and that number - 1.