1: \begin{abstract}

2: It is well documented that 

3: a model for the underlying asset price process 

4: that seeks to capture the behaviour

5: of the market prices of vanilla options

6: needs to exhibit

7: %asset returns can exhibit large fluctuations over very short time intervals.

8: %It is also generally accepted that 

9: both diffusion and jump features.

10: In this paper we assume that the asset price process 

11: $S$

12: is Markov

13: with 

14: c\`adl\`ag

15: paths

16: and propose a scheme for computing

17: the law of the realized variance of the log returns

18: accrued 

19: while the asset was trading in a prespecified corridor.

20: %quadratic variance of the underlying process,

21: We thus obtain an algorithm for pricing and hedging 

22: volatility derivatives and derivatives on the 

23: corridor-realized variance

24: in such a market.

25: The class of models under consideration is large, as it

26: encompasses

27: jump-diffusion and 

28: L\'evy processes.

29: %the stochastic volatility models with jumps

30: %described in~\cite{CarrMadanGemanYor_LocalLevy}.

31: We prove

32: the weak convergence of the scheme 

33: and describe in detail

34: the implementation of the algorithm 

35: in the characteristic cases

36: where

37: $S$

38: is

39: a CEV process

40: (continuous trajectories),

41: a variance gamma process

42: (jumps with independent increments) 

43: or an

44: infinite activity jump-diffusion

45: (discontinuous trajectories with dependent increments).

46: \end{abstract}

47: