The 'Math' and 'Numbers' sections of Phils's book are a good place to start.
I'm assuming that by 'decimal number', you mean the value of 50/3 expressed as 16.6666...
To do it, your PLC needs to be able to do floating-point calculations and the HMI needs to be able to display it properly. There is an IEEE standard method of encoding floating point numbers in 32 bits. The PLC and the HMI need to be instructed to interpret those 32 bits as a floating-point value and not as a 32-bit signed or unsigned integer.
If your PLC is limited to integer math, all is not lost. When you (integer) divide 50 by 3, the result is 16 if your system truncates the result, 17 if it rounds up. But when you (integer) divide 50000 by 3, the result is either 16666 or 16667, and you can instruct the HMI to insert the decimal point in the proper location.
When you do stuff like that with integer math, you have to be careful to anticipate the maximum and minimum values you'll be using. For example, the maximum unsigned 16-bit integer value is 65535. If you needed to divide any value between 25 and 75 by three, in your example of 50/3, you could only use 5000/3 to get two decimal places. If the maximum value of the numerator is 65 or less, you can get three decimal places. If your PLC supports 32-bit integer math, the range if values is +/- 2 billion (two thousand-million to you) and change.