Business models for emission management

This thesis focuses on multistage stochastic programming for CO2 emissions management models. It defines the multistage stochastic programming in general, portfolio selection problem with utility function but also with other new approaches such as using risk measures, chance constraint or second order stochastic dominance. For testing this models we need to also reformulate our problems to fit scenario tree which are generated. For some models we also need to reduce the dimension of scenario tree. Thus some techniques for scenario tree dimensionality reduction are discussed. We try to apply all these approaches to our data and get results on real data from power energy sector. For this sector, the decisions about allowances might be very important as they are not granted any allowances