Essay: THE EVOLUTIONARY PROGRAMMING TECHNIQUE

(10) Presented evolutionary algorithm as efficient techniques used to take care of optimization faults, in which the search is divided into three groups: evolutionary programming, genetics algorithm and evolutionary strategies. These heuristics have been generally utilized because of their relative preferences to other optimization techniques, for example, they don’t require a predefined kind of objective capacity; they are productive in the search for the optimal planning arrangement in moderately big search spaces; and they can be easily adjusted to changes in the problem variables.
As a few others heuristic systems used to determine the combination optimization placement and sizing problems, the evolutionary algorithm do not promise that the best solution set up is the general optimal of the problem; on the other hand, it is very likely that the perfect solution is an acceptable approximation of that optimal. Requiring evolutionary algorithm can be applied to solve continuous problems and other previous discretization. The evolutionary programming is a simulation of the development procedure of a populace of people along various generations. In this population, every individual serve as only possible solution of the optimal placement problem. The best adjusted people will be sustained, as the generation pass by, the best adjusted people will be sustained, i.e., the solution that present the best objective capacity qualities pass to the subsequent generation.
(10) Suggested four stages that ought to be taken after, generally to design an evolutionary programming algorithm:
* Solution codification must be determined;
* Creation of initial population is needed;
* characterize the combination guideline;
* characterize the selection guideline.
The algorithm is iterative and the most utilized, the combination standard produces new individuals from others that are already in existent. The evolutionary programming commonly uses only single operator in the combination procedure and simulation is common. Simulation produces a new individual by an irregular perturbation in the attributes of an arbitrarily chosen individual. The selection guideline is utilized to figure out which individual of a generation will go to the following one and the wellness of an individual is characterized as the objective function used for the solution codification.
Useful paper 2
[105] was firstly introduce EP, and later presented ES and GA.[45]GA is implemented by DRG to determine productively the optimal placement and sizing.GA is different from other conventional search algorithms due to its high capacity searching and optimization technique based on a model of evolution adaptation in nature. GA does not require subsidiaries or other assistant information and it works with a populace of individuals and every individual stands for a solution.GA is assessed due to its solution quality and its fitness, which is estimated by utilizing fitness function.in DRG position according to (108), this fitness is determined based on reduction in real power losses, providing optimal size, to decrease investments and operational costs. From the various test results,[40] explain HRA performance over the conventional method. An enhanced Hereford Ranch Algorithm (HRA) was executed in 1998 with one objective function to reduce active power loss and contrasted with second order techniques and traditional GA. Khatod et al. presented and develop EP based methodology for discovering the optimal placement of photo-voltaic arrays and wind turbine generators in a radial distribution network [51]. Considering the limitations on bus voltages, line loading, number of DRGs to be put and dispatched wind influence then active energy loss will be minimized. Various researchers [42, 44, 46, 48, and 49] actualized GA to maintain single objective, which decrease the total real power loses in the distribution system. , Borges et al. [35], Teng et al. [50], and Shaaban et al. [54] proposed a key DRG position system by executing GA [35, 50, 54]. The fitness assessment function that drives the GA to the solution is derived as an advantage/cost relation, where the advantage is measured from the deduction of electrical losses, and the expense is depending on installation and investment. Borges et al. also utilized GA to estimate the DRG impact on unwavering quality, misfortunes and voltage profile alongside DRG planning [35]. Besides, in 2012, Shaaban et al. [54] seriously thought about the instability and variability connected with the power output of renewable DRG and also the load variability.
GA has been utilized by numerous researchers to handle multi-objective (MO) in DRG situation [34, 37, 38, 47, 53]. In 2005, Celli et al. proposed a MO guideline for the placing and sizing of DRG material into existing distribution system, the methodology characteristics was accomplished by minimizing distinctive functions.
Carpinelli et al. [36] expanded more about the MO approach to incorporate uncertainties in DRG energy generation. [109] explains why every conceivable future is figured as a situation accordingly, a ”double trade off technique” is utilized. By utilizing MO e-constrained procedure in the first trade off, it permits DRG placement and sizing solution for every situations considered, for example, combination of various kind of wind speed in all the possible location; isolating the most robust solution is allowed in the second trade off. Along these lines, the operator is totally allowed to drive the optimization in a certain direction without losing objectivity and simplification [36]. In 2007 Haesenetal [39], Proposed a well explained MO planning technique for the coordination of stochastic generators in distribution systems. This methodology uses the Strength Pareto Evolutionary Algorithm (SPEA) calculation, an antecedent of the SPEA2 calculation [39]. It was revised to the SPEA2 calculation in 2009 by Rodriguez et al., reached out to incorporate the analysis of controllable DER units and upgraded to make note of qualities that reflect environmental effect and voltage quality [34]. Another MO programming methodology in view of Non dominated Sorting Genetic Algorithm (NSGA) was practiced by Ochoa etal.to locate the best configuration that increase the integration of distributed wind power generation while fulfilling voltage and thermal capacity [43]. Kumaretal [41] in 2010 presented the DRG integration approach with MO model was actualized for speedy restoration and to decrease the additional power demand during use of GA under cold load pick up .In 2011, Moeini Aghtaieetal. [52] Implemented NSGA to reduce the aggregate costs, total losses, and enhance framework unwavering quality in the distribution network.
Final paper
2.2.1 Fuzzy Logic
Masoum et al. (2004a) applied fuzzy logic for solving the discrete optimization problem of fixed shunt capacitor placement and sizing under harmonic conditions. Power and energy losses due to installed capacitors and the cost of fixed capacitors are used as the objective function. Kannan (2008) and Saranya (2011) developed fuzzy expert system to determine suitable candidate nodes for determining the optimal capacitor sizes in distribution systems. Bhattacharya et al. (2009) formulated new fuzzy membership functions to identify probable capacitor locations in radial distribution systems. A new algorithm for selecting capacitor nodes was presented, and simulated annealing technique was employed for final sizing of the capacitors.
Useful paper 2
The information and parameters utilized as a part of DRG sitting are normally derived from different sources with a wide variation in accuracy. For instance, load is studied as known and indicated in all techniques, despite having a high instability. Likewise, cost of DRG, electricity market price, peak power saving expense of DRG and so on may be subjected to uncertainty to some degree. Along these lines, questionable binds because of lacking data may create unverifiable locale of choices. Therefore, generation of uncertain region of decision is concerned due to insufficient information. Consequently, the validity of the outcome from average values cannot serve as uncertainty level. To account for the instabilities in the data and objectives identified with different and typically conflicting purpose in DRG arrangement, the utilization of fussy set hypothesis may assume a significant part in choice making [107]. In 2006 Ekel et al, implemented a multi objective allocation of resources utilizing Bellman-Zadeh approach to determine a fussy situation.
Base on the Bellman-Zadeh approach authors developed parallel Adaptive Interactive Decision-Making System (AIDMS) [120]. Lalitha et al. and Kumar et al. proposed fuzzy set hypothesis to determine the suitable location for DRG sitting. Two aims were considered while designing a fuzzy logic for describing the optimal DRG location, which were to reduce the real loss of power and to control voltage within reasonable limits. Voltage stability and loss of power values of the distribution network were designed into voltage stability index (VSI) and power loss index(PLI) to get DRG suitability file (DSI) as result [71,72].
2.2.3 Hybrid Artificial Intelligent Techniques
To design a hybrid intelligent system, two or more AI method is utilized. Within the past decade, hybrid intelligent systems have been used in electrical engineering functions.(Al-Mohammed and Elamin 2003) discuss about a combination optimization problem with a non-differential objective function has been planned and solved utilizing GA, TS, simulating annealing and hybrid GA-fuzzy logic algorithms. According to Hsiao et al. (2004) the capacitor placement problem in distribution system is solved using combined fuzzy GA method. Three particular goals were considered; improve the voltage profile, minimize the total cost of energy loss and capacitor and maximize the margin loading of feeders.(Das 2008) proposed a method where a GA based fuzzy multi objective approach for optimal location while improving voltage profile and increasing net savings in a radial distribution network. This study endeavored to increase net saving and re voltage reduce node voltage deviations.
According to (Ladjavardi et al. 2008) he proposes a GA-fuzzy logic algorithm for determining the discrete optimization problem for placing and sitting in the presence of voltage and current harmonics. Fuzzy logic and particle swarm optimization (PSO) have been applied to solve the optimization location and sizes in an electrical system respectively for reactive power compensation of radial distribution systems. A fuzzy objective capacity with bacterial searching technique was designed to make capable optimal tool for solving optimal placement in an electrical system problem considering both loss deduction and improved voltage (Tabatabaei et al. 2011). Mohkami et al. (2011) presented the optimal placement of DG to utilize bacterial foraging with PSO considering a multi objective function. (Hooshmand et al. 2011) also said that the bacterial foraging PSO procedure was likewise connected to explain the optimal placement of DG to improve voltage profile and minimize cost of energy losses (Hooshmand et al. 2011).
NOT PALGE
Paper 2
4.3. Artificial intelligence (AI)
approaches inspired by evolutionary mechanism such as selection, crossover and mutations [111,112]. They are efficient optimization search techniques employed in finding the exact or near-optimal solutions in multi-objective optimization problems. Applications of AI to complex problems are found in several disciplines such as bioinformatics, computational science, engineering, chemistry, mathematics etc. A genetic search is usually preceded with a randomly generated initial population, covers the whole range of possible solutions, otherwise known as the space. The fitness of each individual in the population in each generation is then evaluated and thereafter modified to form a new population of better solutions. This new populations is then used in the next iteration of the algorithm that terminates either when a satisfactory level of fitness has been attained or when a maximum number of generation have been produced in the population. Of the literature reviewed in this study, some researches [28,57,113,115,116,117,118,119] adopted the GAs based AI approach in finding the optimal size and site of DG units power distribution systems, though Carpinelli et al. [113] did not optimize size in their work
2.2.2 Evolutionary Computation
According to Khajehzadeh et al. 2011 different types of evolutionary computation based optimization techniques have been used to search for optimal or near optimal solutions for optimal capacitor placement in radial distribution networks. These evolutionary computation techniques include Tabu search (TS), simulated annealing (SA), ant colony optimization (ACO), harmony search (HS), genetic algorithm (GA), and particle swarm optimization (PSO). TS is based on the hill-climbing method which evaluates the final solution by repeating the process of creating solution candidates in the neighborhood around the initial solution and selecting the best solution among the candidates (Eslami et al. 2011).
The hill-climbing method stops if the solution is not improved and therefore it can be easily trapped in a local minimum (Mori et al. 2005). Optimal capacitor placement has been solved with a hybrid method that utilizes TS and other heuristic techniques such as genetic algorithm and simulated annealing. (Gallego et al. 2001), Mori et al. (2005) applied a variable neighborhood TS technique for capacitor placement in distribution systems. An optimization technique that exploits the resemblance between a minimization process and crystallization in a physical system is simulated annealing (SA). Ghose et al. (2003) applied SA for solving optimal capacitor placement problem considering unbalance and the presence of harmonics in a distribution system. A modified SA technique has been developed for optimal placement and sizing of fixed capacitor banks in a distribution network (Elmitwally 2011). Another optimization technique that is inspired by the behavior of ants in nature such that the ants can find the shortest path from their home to food is the ant colony algorithm (Eslami et al. 2011). Annaluru et al. (2004) applied the ant colony algorithm to solve the capacitor placement and sizing problem, in which the Newton-Raphson power flow method was used to calculate the cost function. Chang (2008) solved the optimal capacitor placement problem, the optimal feeder reconfiguration problem, and a combination of the two using the ant colony algorithm.
The optimal capacitor placement and sizing is also solved by using genetic
algorithm (GA) which searches for an optimal solution using the principle of evolution based on a certain string which is judged and propagated to form the next generation (Goldberg 1989). A GA-based method that incorporates nonlinear load models for the problem of finding optimal shunt capacitors in distribution systems has been reported by
Abou Ghazala (2003). The optimization problem was formulated such that the optimal solution did not result in severe resonant conditions at harmonic frequencies. Another GA-based method was developed for solving the discrete optimization problem of fixed shunt capacitor placement and sizing in the presence of voltage and current harmonics (Masoum et al. 2004b). The problem of simultaneous placement and sizing of both capacitor banks and voltage regulators in unbalanced distribution system in the presence of linear and nonlinear loads, using GA is discussed in Carpinelli et al. (2006). The problem is formulated as a mixed integer program that accounts for imbalance and incorporates network losses and costs of capacitor, voltage regulator and harmonic distortion. Reddy et al. (2007) applied GA and a power loss index to determine suitable candidate nodes in distribution systems for capacitor installation. The power loss index is used to determine the suitability of capacitor placement at each node. The buses with the highest suitability are identified for capacitor placement. Carpinelli et al. (2009) applied GA for the optimal sizing and setting of capacitors in unbalanced multi-converter distribution systems. Swarnkar et al. (2010) presented an efficient method for determining the optimal number, location, and sizing of fixed and switched shunt capacitors in radial distribution systems using GA.