Solution frequency is a frequency of operation for which antenna is designed when you are analyzing first time. 

Subsequently you may change it if you wish to see the operation in other band.

If you have designed antenna for multi band operation, it could be multiple frequencies. 

Depends upon the solution frequency, size of the finite elements (meshes) will vary in HFSS. Generally, for an UWB antenna highest frequency is recommended as solution frequency. But, as per my experience the result will not differ much if the solution frequency varies.  

As mentioned in HFSS design guide, in case of fast simulation, center frequency is recommended as solution frequency. In case of Discreet, highest frequency is recommended.

Apply sweep in HFSS and then find out resonant frequency of antenna, according to results, alter the antenna dimensions by meandering it and electrically lengthen the antenna to decrease the resonant frequency

if you are solving for a large BW then perhaps you can use a different solution SW, which will solve in the time domaine such as CST and then you do not have to define the frequency but rather an input signal in the time domain.

Choosing the resonant frequency is important issue in the design , choosing it depends on the applications that you used, for example, if you will use the antenna for Bluetooth applications, then the resonance frequency will be different than WI-Fi application.

But I prefer to use the geometric calculations, which it connect the parameters of resonance frequency and dimensions of antenna.

Selecting solution freq. in solution setup of HFSS depends on frequency sweep types such as fast, interpolating, and discrete type. When sweep frequency range is 3 GHz to 10 GHz (UWB), solution frequency should be 6.5 GHz for fast sweep (center frequency), 10 GHz for interpolating (higher frequency). In discrete sweep, any frequency is OK as solution frequency because all sampling frequencies become solution frequency in each mesh anaysis.

Do discrete( maximum number of points)  sweep. Keep the solution frequency as higher frequency.

I have to agree with the majority here, when dealing with large bandwidths I've generally found it best to mesh at (set the solution frequency to) the highest reasonable value.

This will result in the highest density (and therefore, most accurate) mesh, while the fact that you're working with a UWB design means that you should be able to avoid meshing at a frequency where a different, undesired mode is dominant.

UWB is a wideband antenna from 3 GHz to 10GHz.  Questioning here where is the resonance frequency 3Ghz , 10GHz or the middle  (6.5GHz) . I think 6.5GHz is ok . 

I want to add an important point, when you chose the resonant frequency, you must change the dimensions of the antenna to get a good results in term of return loss and bandwidth 

solution frequency should be 6.5 GHz, the center frequency of UWB. While you set up the simulation, not go for interpolating, go for fast simulation, enabling "save the fields for all frequency" and after that observe the frequency vs gain plot.

If the designed antenna generate positive gain for all the frequencies (3 GHz-10 GHz), then it can be concluded that it is working properly through out the band.

Thank you all for your suggestive reply.

First u solve mathematically and try to find the resonant frequency of your structure for particular mode. Like find resonant frequency for dominant mode and then same frequency should be kept as solution frequency in simulation. Now using this procedure u can plot the fields and u can also define same mode for same solution frequency. 

If your design is new which is not easy to analyse mathematically...then you can assume any of the frequency between selected frequency sweep. In such cacses most probably deepest point in S11 plot shows resonant frequency for any mode which exists in the structure.

Solution frequency = designed frequency

But what of if am using CST not HFSS what should i put or are they the same thing?