Our motivation is lying on the questions "How does a food scare and information of a food scare influence the buying decision of one single agent and the changes in the aggregate demand?" and "How can we evaluate effective risk communication strategies?". Since we investigate a society of consumers which do have their own decision functions, we can observe how new information could influence the behaviour of each consumer and more interesting the aggregate changes in the demand by creating a population of agents. This multi agent simulation can be used to investigate how different information releases and decision functions influence the aggregate demand. The agents get information from the networks in which they are present. Then this information will be processed, the trust regarding the food item under investigation will be updated and taken into account for the own decision. Our intention is to measure how different risk communication strategies influence the aggregate demand. For this purpose we use a multi agent method in order to follow a bottom up approach where each agent acts individually. The interaction between the agents leads to an emergence of an aggregate demand that comes from the bottom up. Each agent follows its internal updating and decision algorithms so that on the aggregate level the demand changes according to the outcomes of the interaction and its related updating processes. After the communication phases the aggregation of the outcomes of each agent shows the result of the information strategy that was selected. In this way we can test and investigate different risk communication strategies and evaluate these information policies, i.e. the benefits of a risk communication strategy can be evaluated with this multi agent approach.