The only tricky thing here is a comparison with tolerance, since because of round up errors you can well never meet
answer == value
condition. The implementation could be
double answer = 300.0;
double tolerance = 1e-10;
while (true) {
// based on previous answer we compute next one
double value = 0.5 * (answer + 123.0 / answer);
//TODO: you can print out steps here, if you want something like this
//Console.WriteLine(value);
// check convergence with tolerance
if (Math.Abs(answer - value) <= tolerance) {
answer = value;
break;
}
// next answer (value) becomes the previous one (answer)
answer = value;
}
// 11.0905365064094
Console.Write(answer);
The actual answer (prove it) is just a square root:
// 11.09053650640941716205160010261...
Console.Write(Math.Sqrt(123));
Real life implementation (if my boss wants me to implement it):
public static double NewtonEstimation(Func<double, double> function,
double tolerance = 1e-10,
double guess = 1.0) {
if (null == function)
throw new ArgumentNullException("function");
else if (tolerance < 0)
throw new ArgumentOutOfRangeException("tolerance", "tolerance must not be negative");
while (true) {
double value = function(guess);
if (Math.Abs(value - guess) <= tolerance)
return value;
guess = value;
}
}
...
// 11.0905365064094
Console.Write(NewtonEstimation(x => 0.5 * (x + 123 / x)));
9
solved C# logic (solution needed in coding) [closed]