Question about using Linear regression (Y=a+bx) in cloudsim plus
(VmAllocationPolicyMigrationLocalRegression.java)
As input y=utilization history [10]
And x =[1,2,3,4,5,6,7,8,9,10]
Compute output
Hostutilization= estimates[0] + estimates[1] * (length + migrationIntervals)
Y1 host utilization
A= estimates[0]
B= estimates[1]
But I don’t understand x1 is it will be constant value
X1 (length+migrationinterval)
migrationIntervals = Math.ceil(getMaximumVmMigrationTime(host) / getSchedulingInterval());
protected double getMaximumVmMigrationTime(final Host host) {
//@TODO It must compute the migration time based on the current RAM usage, not the capacity.
final double maxRam = host.getVmList().stream()
.map(Vm::getRam)
.mapToDouble(Resource::getCapacity).max().orElse(0);
return maxRam / (host.getBw().getCapacity() / (2.0 * 8.0));
}