Genetic algorithm refers to the method that is used in solving optimization problems that are either constrained or unconstrained. The theory of genetic algorithm bases its argument on the process of natural selection and offers a clear mimicry of biological evolution. The population of individual solutions is continuously modified by the genetic programing algorithm. In some cases, genetic algorithm is known as the evolved antenna model. Heuristics that mimics the natural selection process in artificial intelligence is also searched with the help of genetic algorithm (David 1996):. These heuristics is usually used to generate solutions to a variety of optimization problems.
According to genetic algorithm, individuals from current population are randomly selected at each step and used in producing children for the next generation. The population evolves towards a solution that is optimal. In cases where we have problems that are not easily solvable by the various standard optimization algorithms, we prefer the use of genetic algorithm. These problems could include stochastic processes, nonlinear, non-differentiable and continuous functions.
Genetic algorithm is a unique. It differs from classic algorithm in different ways. Firstly, genetic algorithm generates a population of points while classical algorithm generates just one single point for each iteration. In genetic algorithm, we take the best point to be that point that approaches the optimal solution. The classical algorithm has its sequence of points approaching the optimal solution. Deterministic computation is used to select the next point in the sequence while the genetic algorithm usually generates random numbers and uses the selects the random numbers selected in the respective computations.
Variety of optimization problems can be solved using genetic algorithm. Each candidate usually has properties traditionally altered. The solutions obtained usually are represented binary as character strings. They are represented in terms of 0s and 1s that require serious encoding before interpretation. We usually refer to the individuals produced at each iteration as generations. The individuals of all generations are evaluated for fitness according to the genetic programing algorithm. A solution to the objective function usually gives us the optimal solution to the stated problem.
Well placement applications within Genetic Algorithm
This research study discusses how Genetic algorithm uses the optimization of well placement. We developed a simple genetic algorithm program to optimize well placement. We analyzed the effects on the performance of the algorithm on different internal parameters. We as well suggested a reliable base configuration and the most preferable recommendations for further research and study.
It is every businesss key objective to ensure that they yield maximum profits from their operations. Even for the oil field, proper strategies should be put in place to ensure that the oil fields yield maximum profits. As observed, it is usually found that the obtaining maximum profits from oil field investments require a lot of sacrifice and dedication. The frontiers that are currently being produced in most cases are very costly. This is due to the current scarcity of the easy fields worldwide.
The reservoir performance need to be fully optimized so as to ensure that maximum profits are realized and that newer and better tool are developed. The reservoir engineer should therefore ensure the reservoir performance is as optimal as possible. It is undisputable that optimization of an oil field is an extremely complex procedure. It therefore needs a serious application if genetic algorithm. The procedure involves a lot of variables under consideration. The variables that are most likely involved in the optimization of the reservoir performance includes, reservoir architecture which is a geological variable, well number, platform position and type of platform which are production variables. The oil and gas prices are also monetary variables that need to be put into consideration in any case where optimization of oil field profits is required. It is usually very difficult to determine the objective function and restrictions involved due to the large number of variables. The geological uncertainty of the reservoir is another factor that contributes to this serious difficulty in determining the objective function and the restrictions
In evaluating the objective function, we usually require numerical simulators. The traditional function optimization is possible due to the fact that the reservoir engineers usually lack adequate analytical solutions to the optimization problems. The traditional methods for instance, simplex, Jefferys, gradient and Goldberg usually dont work due to the con-continuity and non-linearity of oil filed optimization and well placement. It is the responsibility of the reservoir engineer to assess the behavior of the reservoir when exposed to different conditions or scenarios with the help of multiphase flow simulators. However, it usually takes a lot of time to put up the possible alternatives and set up each scenario as expected. The length of time and the exorbitant costs usually discourage the engineers from making these decisions. They therefore resort to genetic algorithm for as the best option towards these scenarios. There are a number of artificial intelligence procedures that have been invented to ensure that the process remains as automated as possible. These include the genetic algorithm and the Neural Networks. Just to take a brief overview of the building concept of the genetic algorithm. We find that the Darwins theory of survival for the fittest is used in genetic algorithm to determine the optimal solution.
Genetic Algorithms as earlier defined refer to the algorithms that use natural selection mechanics to for optimal solutions to variety of problems. The main property of genetic algorithms is the natural selection technique that it applies in its application. The natural selection technique is in some cases referred to as the survival for the fittest mechanism. Genetic programing in the case of well placement is a purely non- random process. This is due to the fact that the process combines the survival natural selection approach with the stochastic or random exchange of necessary information. The main reason why the process does this is to ensure that the efficiency in well placement and other operations is deemed achievable. The scientists used these methods in the early 1960s in simulating evolution process by studying the mutation patterns. John Holland was the first scientist to try using genetic algorithm in the early 1970s in solving complex problems. He developed the basic or the skeleton theory and demonstrated the ability to represent complex problems using various chains of algorithms. The algorithm chains that he developed were further improved by a chain of simple transformations. John Holland attempted to show that there existed the possibility of finding the optimal individual by evaluating only on a very small population. He found that there were 900 individuals in his population of study which consisted of 2.71011people.
There are four major differences between the genetic algorithm and the methods that were traditionally used in solving complicated scenarios such as well placement in the oil fields. They include direct manipulation of coding procedures, full population search rather than point-wise searches, blind searches characterized by random sampling procedures and using stochastics operators to aid in the search. Genetic programing successfully overcomes the challenges and limitations in its applications especially well placement due to these properties. It is more recommended than the traditional methods due to its use of objective function in a more continuous and objective function.
Genetic algorithms structure
In this section, we analyze the different components of a genetic algorithm. We as well analyze the variations in detail. According to genetic algorithm, individuals from current population are randomly selected at each step and used in producing children for the next generation. The population evolves towards a solution that is optimal. In cases where we have problems that are not easily solvable by the various standard optimization algorithms, we prefer the use of genetic algorithm. These problems could include stochastic processes, nonlinear, non-differentiable and continuous functions.
Genetic algorithm is a unique. It differs from classic algorithm in different ways. Firstly, genetic algorithm generates a population of points while classical algorithm generates just one single point. In genetic algorithm, we take the best point to be that point that approaches the optimal solution. The classical algorithm has its sequence of points approaching the optimal solution. Deterministic computation is used to select the next point in the sequence while the genetic algorithm usually generates random numbers and uses the selects the random numbers selected in the respective computations
This can be classified as the first step in the process of algorithm development and problem solving. In this case, the problem variables involved are represented as chromosomes by coding. The resultant chromosome from the initial coding is generated randomly or by using algorithms. We are assured of the appropriate population coverage. This is referred to as fangalgorithm. The introduction of part of the population can be manually be implemented by the programmer. The interpretation will be based on whether or not the optimum result is needed. In terms of size, chromosomes are usually made up of information bytes that usually represent the information previously codified in the problem. The problem variables involved in the coding differ in size from depending on the size of information it may contain. A binary code or gray is usually used in codifying every variable. The codification in most cases is a decimal before simplification. We shall further discuss the criterion we need to use in simplifying the algorithm in details.
In this step, the programmer or the well engineer evaluates the functionality and reliability of all chromosomes. The evaluation protocol if not properly checked can prove more complicated than required. The evaluation procedures in most cases have technical and economical functionalities fully incorporated. The chromosomes are ranked from the worst to the best under this evaluation section. Biologically speaking, every chromosomes fitness to survival and reproduction by every chromosome is usually represented during the evaluation period. It is usually very preferable not to use extremely strong penalties in solving complex evaluations. These penalties may restrict the populations evolution. It is the responsibility of the reservoir engineer to assess the behavior of the reservoir when exposed to different conditions or scenarios with the help of multiphase flow simulators. However, it usually takes a lot of time to put up the possible alternatives and set up each scenario as expected. The length of time and the exorbitant costs usually discourage the engineers from making these decisions. They therefore resort to genetic algorithm for as the best option towards these scenarios
This usually represents the most involving step in the process the design of genetic algorithm in the process of well placement. It is usually recommended that we chose the chromosome with the highest number of variations since it can be more adaptable and hence has the hi...
Cite this page
Check the Brief Overview of Genetic Algorithm in a Free Essay. (2019, Dec 02). Retrieved from https://speedypaper.com/essays/brief-overview-of-genetic-algorithm
If you are the original author of this essay and no longer wish to have it published on the SpeedyPaper website, please click below to request its removal:
- Problems in Usage of FIOA - Essay Sample for You
- Relationships Research Methodology, Data Analysis Essay Sample
- Leaving a Legacy - Free Essay Example
- Political Parties: Party Politics at The Grassroots, Essay Example
- Free Essay on the Methodology of Maximizing the Profit Margin of MEP Projects
- Comparative Ethnic Experience: Chinese-American and Hispanics, Italian-American
- The Importance of Math for the College Students, Free Essay for You