In this paper a novel non-parametric approach for designing geodetic networks based on  evolutionary ALGORITHMS is proposed. Since conventional analytical optimization methods are based on gradient approaches, they have some drawbacks including approximate parameters requirement and inability to work with discrete objective functions and/or constraints e.g. visibility matter. To prevent from above problems, the standard GENETIC algorithm method is utilized. Primary tests demonstrate the capability of proposed method to design the optimal network without approximate values and with visibility constraint satisfaction.    

There are some machineries and equipments as well as relatively fix amount of human resources involved in any production process. In many cases, by observing operational method of a process, one may see that some of these machineries are not working, but on the other hand, other equipments are continuously working, and many parts are set in front of them, waiting to be set up over machines to continue production procedure. Existence of non-working time as well as overload tasks, that means imbalance in production process, are among factors creating some problems for system’s management. One guideline that a manager may use to encounter imbalance problem of production process is assembly line balancing. In this thesis, GENETIC algorithm method is used to solve assembly line balance problem. This algorithm presents a new method during execution of crossover operation and combination of parent chromosomes in order to produce child and mutation building in chromosomes. Finally, efficiency of obtained answers by GENETIC algorithm is compared with numeric method. 

Each attack graph represents a collection of possible attack scenarios in a computer network. In this paper, we use weighted attack graphs (WAGs) for vulnerability assessment of computer networks. In these directed graphs, a weight is assigned to each exploit by the security analyst. The weight of an exploit is proportionate to the cost required to prevent that exploit. The aim of analyzing a weighted attack graph is to find a critical set of exploits such that the sum of their weights is minimum and by preventing them no attack scenario is possible. In this paper, we propose a greedy algorithm, a GENETIC algorithm with a greedy mutation operator, and a GENETIC algorithm with a dynamic fitness function for analyzing the weighted attack graphs. The proposed ALGORITHMS are used to analyze a sample weighted attack graph and several randomly generated large-scale weighted attack graphs. The results of experiments show that the proposed GENETIC ALGORITHMS outperform the greedy algorithm and find a critical set of exploits with less total weight. Finally, we compare the performance of the second GENETIC algorithm with an approximation algorithm for analyzing several randomly generated large-scale simple attack graphs. The results of experiments show that our proposed GENETIC algorithm has better performance than the approximation algorithm and finds a critical set of exploits with less cardinality.