Metapopulations are a popular way to gain an abstract model of a certain species in a scattered landscape. These models can be used for analytical questions concerning the viability of the modeled species. Several applications exist that evaluate metapopulations concerning their persistence, e.g. Meta-X. There is a strong interest in discovering modifications of the environment, that have a positive influence in the survivability of the population. Recommendations for those modifications, that can be evaluated by tools like Meta-X, are mainly given from experts. Automatically generated recommendations were hardly available mainly because of the high complexity of the problem. This paper shows the results of an implementation of an optimization algorithm for methods to increase viability of metapopulations. It uses a metamodel of a metapopulation model for accomplishing the needed performance.