The energy costs of melting snow are huge. It's done frequently for small areas (private driveways, parking lots) where cost efficiency isn't a concern (the wealthy) or where heat energy costs are very low (geothermal sources, i.e. Iceland).
To do it on a wide scale would require ridiculous amounts of energy. I guarantee that nobody would be prepared to foot the bill for that, and these solar panels are highly unlikely to produce even a tiny fraction of the amount of energy required, unless you amortize their production over a longer period and store it somehow.
Unfortunately, it's not just a "little more costly", it's massively more costly, to the point where it's just not a good idea.
http://www.concretenetwork.com/concrete/snow_melting_systems...
There doesn't need to be sufficient yearly yield. This isn't just trying to pay for itself by collecting electricity on clear days to melt snow. It'd still be a great solution if it uses far more electricity than it collects because it's competing with snow removal trucks. If it could be cheaper and more efficient than having trucks driving around dropping salt and plowing, it'd be amazing.
Even if it's a little more costly, the ability to keep all the roads clear, all the time, would be incredible and worth some added expense. Think about how much safer the roads would be if they could remain free of snow.