摘要
The optimal allocation of regulators banks in distribution systems is a merely combinatorial problem in which the best points of installation correspond to the best benefit, considering the admitted objective function, without violating and operating limits. The objective function must be chosen so that its value represents the operation state of the system. As the problem possesses combinatorial nature, its complexity will increase exponentially with the number of possibilities. Systems with large numbers of nodes and / or with the possibility of installing more than one bank require a large number of calculations to find the solution. An additional issue is the fact that the problem does not have a continuous nature, presenting discontinuity points in the objective function, limiting the application of optimization methods based on gradients. Based on the nature of the problem two optimization methods were used to solve the problem: Genetic Algorithm (GA) and modified Tabu Search (TS). The GA function will scour the search space and find regions with local minima that are candidates to be the solution. On the other hand the TS provides local search in the regions defined by GA so that the overall optimum is achieved.
The optimal allocation of regulators banks in distribution systems is a merely combinatorial problem in which the best points of installation correspond to the best benefit, considering the admitted objective function, without violating and operating limits. The objective function must be chosen so that its value represents the operation state of the system. As the problem possesses combinatorial nature, its complexity will increase exponentially with the number of possibilities. Systems with large numbers of nodes and / or with the possibility of installing more than one bank require a large number of calculations to find the solution. An additional issue is the fact that the problem does not have a continuous nature, presenting discontinuity points in the objective function, limiting the application of optimization methods based on gradients. Based on the nature of the problem two optimization methods were used to solve the problem: Genetic Algorithm (GA) and modified Tabu Search (TS). The GA function will scour the search space and find regions with local minima that are candidates to be the solution. On the other hand the TS provides local search in the regions defined by GA so that the overall optimum is achieved.
