I do not have a specific paper to explain step-by-step the procedure, but I would suggest to check yourself certain features of the previous tests.

Here, I assume that the "optimum" shut-in time is the shortest time that provides you with most of (or all) the data you need for your analysis.

If the bottom-hole shut-in pressure (BHSIP) of previous tests levelled out on a pressure value that you know or calculate to be the real formation pressure at the time of testing, then the shut-in period was way too long. You can cut the shut-in time to a time that is sufficient to record enough pressure points to correctly extrapolate the real formation pressure at the time of testing (try modelling using sub-sets of the previous data). Too long shut-in times are usually selected when the actual formation permeability is much larger that the formation permeability anticipated before the test. This is a relatively rare case, though.

On the other hand, if the formation permeability is lower or much lower than the anticipated permeability before the test, the selected shut-in times are usually too short. This may happen also when ongoing production causes unforecasted reservoir compaction, which reduces porosity (slightly) and permeability (sometimes considerably, particularly when the reservoir is poorly cemented sand and it is originally overpressured relative to normal hydrostatic gradient). Production-related subsidence can be a good indication of ongoing reservoir compaction and associated reservoir permeability reduction.

Another case of too short shut-in times could be when the reservoir has poor matrix permeability and high fracture permeability. In this case, especially if the reservoir is carbonate and has open vugs connected to the natural fracture system, suitably long shut-in times are requested to correctly assess the vugs storativity and also the matrix storativity.

 Design based on the low values for the parameters mentioned by Roberto. Also, is the aim of the test same as the previous test's aim? For example, previous test might have had an aim of determining the permeability and the new test has an aim of determining the boundaries' distances. I suggest you have a look at the manual of saphir. Also, be careful with the temperature since it affects the compressibility of the gas highly and it can impact the storage period length (in case of tubing head shut-in). Good luck. 

We have Dbranes differential String Theoretically analysed that Quantum Wells continuum mechanics are best manged by diameter-depth ratio  which gives a measure of the field. So if you consider the economic value of the gas field flow by this method over the shut in time and compare it to the  cost of outsourcing the repair you can compute the minimum shut in time by this Econophysics String Field Matching method. See for e.g. Mallick (2016) on this site.

SKM,NH, SM

In addition to the comments above I suggest you use well testing software (such as Saphir mentioned by Kahldoon) to design the well test by running various scenarios of formation permeability, testing time etc. and reviewing the simulated well test response.  Will the simulated response show you the features you are looking for (depends on the purpose of the test) for the testing you have chosen?  And will that happen for all reasonable reservoir and fluid properties?

This might not help but in a real case this is solved by using surface reading of the downhole tool. You can analyze data in real time and stop the test at the optimum time. 

Look up the term "horner plot". According to this procedure you plot the pressure against the time. Time scale is logarithmic. Thereby, even after relatively short times you can estimate the pressure that the reservoir will finally reach.    

Dear Arif the duration of shut-in time depends upon the objectives.  Why we are conducting  a well-test on an already a producing well.  Sometimes just to get some new pressure point to revise the reserves and sometimes detailed reservoir characterization for further development and management.  However, particularly for gas wells, read in details the Isochronal and Modified Isochronal tests.   This is available in any well-test book and monograph.