I'd suggest for white to stick the rook on d7, maybe the bishop on h3 if the black king attacks the rook. Then march the king up on the queenside, maybe pushing the a-pawn as well. It seems like it should be a win.
Only problem is if black can free the rook by pushing the g/h pawn a couple spaces.
that is an interesting position, it looks so easy at first but to find something concrete is really really tricky. at least for an amateur like me
I know that in another one of your threads, you said you didn't see many positions interesting to you at this point in time. I think it was in the smith-morra gambit thread. It occurred to me that perhaps some weaker players had interestign positions in their games, although I assume you'd groan at the bad decisions made throughout the game, but perhaps you'd find an interesting position from a novice's game.
Back to the position, though, is it possible for white to activate the king and immobilize black's king then just go on a pawn grabbing rampage before promoting a pawn of white's own? 
Black wants to play ...g6, ...Kf8 and ...Bh6. If ...Bh6 ? without ...Kf8 first, ...Rd7 forces the bishop to go back to defend d6.
So I think the first thing to do is Bh3 to prevent ...Kf8. If Black manages to bring his bishop to c5 (via h6-e3), even at the cost of the b7 pawn (but not at the cost of the d6 pawn !), he has very very probably the draw.
Afterwards, march your king somewhere and hope... g5 looks juicy, but does it really work ? That's the question.
Hmm, maybe white should place the rook on d8. That would freeze both the rook and the bishop. Bg7 lets d6 get taken. Bh3 means the rook would be unable to be forced away.
I would just take it slow and march my King up the board to sit on b5 or even take b6. I would then march my a-pawn up the board to take care of the extra riff-raff surrounding the King. To my eyes, this produces a threat. It gets the King into the action, opens up files for your light squared bishop and threatens mate or a terrible King position which can be easily exploited.
Then again this is just a minute idea and perhaps a bit too slow.
phylar but that frees blacks pieces i believe, i also had the idea as falcogrine but to my shame i could see no win...
i definetly have to work on my basics technique of converting an advantage, shame no1 seems to talk about that, they always assume you automatically know all that stuff....
also you cant really force black to give up the pawn or maybe you can but its beyond my horizon to calculate that.
I think the right plan is to prevent black from activating his bishop. This will also keep the rook trapped on the back rank to defend it. I would start with g4, intending g5. If h6, then he takes the bishop's square on h6. I would meet that with h4. The idea is to get the light-squared bishop on e6 to defend the rook and get the king to e8 after taking the pawn on b6. If the black king ever tries to get to a7, Bc8 will punish him. Just imagine a position with the white king on b5, the white bishop on e6 and the white pawns on g5 and h5. That would put black in Zugzwang. For black to activate his bishop he needs to push his pawns. Then white can mimic him. If g6, then g4, threating g5. If h6, to stop g5, then h4 will allow white to lock the pawns after either pawn moves. The problem is black will have a weak bishop, that requires the rook to passively defend it. White can improve his position easily and the light-squared bishop and king can already generate threats.
Basically, all of you are finding the difficulties of getting something concrete here.
If there is a FORCED WIN, then a strong engine will find it...Engines are useless in strategical positions.
gaincity, I'd recommend 27. Rd8+.
otherwise, I completely agree. nice job
This isn't that hard of a position as long as you keep in mind the functions of each piece.. black's bishop has to defend d6.. black's rook has to defend the bishop.. and if the black king comes over to defend the black bishop then the black b pawns fall to some combination of the white rook and king.. it'll take a number of moves.. but unless white goes crazy this should be a fairly easy win..
I tried playing against the engine. And two ideas came out. From Gaincity's analysis:
Setup) 25...Bh6 26. Rd7 Bf8 just wasted a move. So instead, 25...g4 fixing the kingside pawns for later. Now:
1) If after that 26. Kb3, 26...Ka7 with the idea of bringing the king to b6 after ...b5. 26...Ka7 27. Bc8 Be7! saccing a pawn for a playable (?) rook ending. This is why 25...g4 fixing the pawns was suggested.
2) And if instead 26. Rd7 b5! since the bishop can't go to d7 now. ...Ka7-b6 follows and it's even harder for white to progress.
I was very suprised when the engine pulled this on me. Tactical resources to reduce to a different endgame.
Setup) 25...Bh6 26. Rd7 Bf8 just wasted a move. So instead, 25...g4 fixing the kingside pawns for later. Now:
I suppose that Bh6 should have been in brackets
This is in response to Remellion's suggestion, allthough there may not be an immediate win, white will have a significant advantage in the endgame
EDIT: the sequence that I posted was set on move 39 which was an accident, I am still new to posting games in this format, if you wish to view it then you should reset to move 26 which is where Remellion suggested me a different move.
Hm... 31...Rh7 trading bishops instead seems better, even though the bishop endgame should be drawn too. Let's see.
To be frank, I have no clue. It looks drawn, in line with "opposite coloured bishops draw".
hmm... in remellion's last line, 26. a4 looks worth looking into.
Another refutation attempt towards Remellion, I still think white can win even with strong black play. Your bishop trade idea was interesting and in my opinion the only line that doesn't lead to a obvious white win, it took me longer to refute that than the BvB endings.
EDIT: Near the end of my mainline I screwed up the notation. Ignore 44...h4 and treat the blue sideline as the main line that begins with 44...Bh6
The following position should be a forced win imo. White's rook on the 7th is super powerful and black can not develop properly, yet the pawn structure is not so easy to exploit.
It's opposite colored bishops so generally speaking, trading rooks will give black one of the easiest draws ever. White must avoid that.
I have analyzed this position to no end, and seemingly white can plays this position for years, and black has to hope he doesn't run out of moves fast otherwise it's lost.
I don't think the engines help in this position. Can you find a forced win? Please be nice about it, if it's obvious to you (which it isn't surely).