Chapter 1 is dedicated to the peculiarities of power company regime control at the conditions of an electricity market. The organisational principles of an exchange have been reviewed, along with the formulation of an energy supply/demand optimisation task at the conditions of uncertainty. A number of methods have been described for addressing the task of energy company regime control. The market-limiting factors that are characteristic of the Baltic countries have been reviewed. Chapter 2 is dedicated to the role and methods of forecasting of the processes that influence the power system. The criteria for computing the inaccuracy of the forecasting procedure have been analysed; much attention has been paid to interrelations of processes and correlation analysis. Chapter 3 contains correlation analysis of the possible input data and the set of input data for further exchange price forecasting has been selected. The author proposes three realisations of the algorithm with artificial neural networks for forecasting energy prices. The chapter forecasts one of the parameters that influence the Latvian power system, namely, the water inflow in the Daugava. Investigation of district heating load forecasting has received detailed and all-round description in Chapter 4. The forecasting of the discussed parameters has been implemented both by ANNs and polynomial models. The problem of input data incompleteness or distortion has been fundamentally investigated. Economic analysis makes it possible to show the costs of district heating load forecasting inaccuracy. Chapter 5 contains optimisation by using the operating regimes of combined heat and power plants and the Daugava hydropower plants. The influence of district heating load, energy price and inflow forecasting on the Latvian power system is illustrated. The chapter contains investigation regarding the operating regimes of the Daugava cascade and the possibilities to maintain a capacity reserve. The costs of maintaining the reserve have been calculated.