We can do the same for a real variable X. In this case X has values within a continuous set. This string length l in this case can be computed according to the inequality. Again, we have 2 to the power l greater equal than this ratio. This ratio has at the denominator the difference between the upper bound and lower bound of the variable and at the denominator the desired accuracy. The presented codings are usable for a single design variable problem. Generally we deal with multi-variables problem and so we have to generate a multi-variable coding simply adding several coding of length l for every design variable.

---