New methods of cooperation between metaheuristics and other algorithmsSometime it is needed to advance the performance of a specific algorithm to solvecertain problem by using metaheuristic algorithms. The complexity of the problem shapes how these programs cooperates in order to solve a specific problem. The objectiveis to make the algorithm adapt to the problem at hand in order to obtain theglobal optimum solution. It is thought that the self-adaptation is improved if cooperationand competition between algorithms are involved. There are several typesof cooperation between GA and other algorithms such as Fuzzy logic , ANN andFuzzy C-Means . But, there are few papers which covers the segmentation of satelliteimages . As an example a promising paper 48 which segments Landsat 8 imageusing semi-supervised method based on GA and trained Radial Basis FunctionNeural Network (RBFNN). In a research paper 49 the authors used multi kerneFCM, ANN, Fuzzy logic , and GA to segment a satellite image to extract someurban features such as buildings and roads. According to the authors their methodwas able to extract roads with an accuracy close to 89 % compared to 80 % foranother which combines only MFCM and ANN only. Another research paper 50uses GA to optimize the weights for an ANN supervised Multi-Layer Percepteron(MLP) 51 algorithm, in order to extract clouds from a weather satellite image.The results of GA-MLP showed better accuracy compared to the results of MLPalgorithm. Fuzzy logic and GA cooperation has played important role in advancingthe satellite images segmentation process, but its use is still limited to adjusting theprobabilities of the reproduction operators for GA during the segmentation processof satellite images . Sumer and Turker 52 used Fuzzy logic to adjust the probabilitiesof crossover and mutation during the segmentation process of high resolutionimage by GA. The method proved to be efficient with kappa index that approached0.88. In this section, unsupervised nonparametric metaheuristic algorithm cooperationwith another two non-metaheuristic algorithms to segment satellite images aredemonstrated using two different examples. There are several reasons for selectingthese two examples such as lacks of papers which cover this area of researchand to prove that these type of algorithms can solve many of the problems whichwere listed in the previous sections (accuracy and speed). This includes the cooperationof Hybrid Dynamic GA (HyDyGA) with Fuzzy C-Means (FCM) 53 heremetaheuristic process role is to improve the performance of FCM in image segmentation. On the other hand, another process which is an ANN algorithm that is calledSelf-Organizing Maps (SOMs) 54 is used to provide the metaheuristic process GAwith initial cluster centers to start from an advanced point in the space of satellitesegmentation solutions.