If you have only a UV detector, you will find some technique to derivatize the analytes due to its very low absorbance and consequently poor sensitivity.
You can get your UV detector working more like a RID at 195-210nm as Robin Joshi pointed out. Using an amino column is one option but dpending on the age of the column, batch and make you may have to fiddle around with the ACN/water ratio to get the best separation. Amino columns can be a bit fickle that way so start at 80/20 but also try 75/25 and 70/30 and inbetween. GOS separation will be done by 20 minutes. Now the ratio is important to whether you want to separate on class (di, tri, tetra etc though an amino column isn't much good out at tetra and larger) or individual oligosaccharide (or disaccharide) species. The more water the faster the elution and less separation of individual oligos/disaccharides
Another option is the use of water s the eleuent and a 8% calcium co-ordinated cation exchange resin such as in the Aminex HPX-87C or the RCM monosaccharide column from Phenomenex (Waters have a column too). Other ion-cordinated resins eg H, Ag,Pb can also be used. One point is that you will have to cation exchange SPE your samples before using this approach as not to change the ion co-ordinated to the resin and alter the elution pattern over time. Sometimes this is refered to as partition chromatography. Again run time is somewher in the order of 20 min but you will need a column oven and the flow will be 0.4-0.6 ml/min isocratic
With ACN/water mixes monosaccharides elute first and with the second set of columns and water last.
Guido is very correct with respect to sensitivity of this approach and derivatisation will help with that issue but generally GOS analysis has plenty of sample to work from unless you are analysing food stuffs where it has been added like pixie dust. The issue with the amino columns is the larger oligos are in smaller amounts and sensitivity makes their quantification difficult. Derivatisation opens up a whole new world of separation possibilities
The other issue you should be aware of is the co-elution of lactose with other disaccharides in both approaches above. Even the use of HPAEC-PAD allolactose and lactose have issues of co-elution. If you define GOS as trisaccharides and larger this won't be an issue.
Unfortunately SAGARS and LC-UV detector are anemies. If you want avoid stomach pains, flee desperately from the UV-detector and become happy with a RID. Normaly both prices use to be compatible. Any case, Paul's proposal works well. Nice elutions!!
After solving the problem of UV detection, you can dedicate to find you own suitable chromatographic separation system. Especially if the selected option is pre-column derivatization. That's another problem, as Paul say...
Best column in my opinion is the NH2 column from Asahipak. Long lifetime good reproducibility. Run 75% ACN 25% water. Works like a charm, but I agree with mg colleagues and say, use RID.