Distributed energy sources are becoming more and more important in the German electricitynetwork. One solution to manage this growing number of distributed assets liesin the Virtual Power Plant concept. A Virtual Power Plant aggregates decentralizedgenerators and loads to behave like a large power plant. Based on new technologies, ituses advanced communication technologies to provide different services (generation ofenergy, steering of power systems, balancing services ...). This thesis proposes a mixedintegerstochastic model of a Virtual Power Plant. The participation to different productsis being studied: selling power on day-ahead basis on the spot market, providingflexibility to the secondary and tertiary reserve market. The particularity of this modelis to consider the revenue generated by the stochastic activation of the reserve market.An operational tool named AlocaBid is implemented in Python, based on the developedmathematical model. The performance of the model is being evaluated using four studycases, representing typical market situations. The results demonstrate the advantage ofthe proposed model over state-of-the-art method for bids’ allocation. / Distribuerade energik¨allor blir allt viktigare i det tyska eln¨atverket. En l¨osning f¨or atthantera det v¨axande antalet distribuerade tillg°angar ¨ar Virtual Power Plant-konceptet.Ett virtuellt kraftverk styr decentraliserade generatorer och laster f¨or att efterlikna ettnormalt kraftverk. Baserat p°a ny teknik anv¨ander det avancerad kommunikationsteknikf¨or att tillhandah°alla olika tj¨anster (generering av energi, styrning av kraftsystem, balanseringstj¨anster ...). Denna avhandling f¨oresl°ar en stokastisk blanda heltalsmodell avett virtuellt kraftverk. Deltagandet i olika produkter studeras: F¨ors¨aljning av maktp°a daglig basis p°a spotmarknaden, vilket ger flexibilitet till den sekund¨ara och terti¨arareservmarknaden. Det speciella med denna modell ¨ar att den tar h¨ansyn till de int¨aktersom genereras av den stokastiska aktiveringen av reservmarknaden. Ett operationsverktygmed namnet AlocaBid implementeras i Python, baserat p°a den utvecklade matematiskamodellen. Modellens prestanda utv¨arderas med fyra studiefall, som representerartypiska marknadssituationer. Resultaten visar f¨ordelen med den f¨oreslagna modelleframf¨or den senaste tekniken f¨or budgivningens f¨ordelning.