Genetic Programming Application Research Articles

Genetic programming application in predicting fluid loss severity

Numerous wells worldwide encounter significant, costly, and time-consuming lost circulation issues during drilling or while deploying tubulars across naturally fractured or induced fractured formations. This can potentially lead to formation damage, wellbore instability, and even blowouts. Effectively addressing this problem and restoring fluid circulation becomes crucial to curbing non-productive time and overall operational expenses. Although numerous methods have been introduced, a universally accepted industry solution for predicting lost circulation remains absent due to the complex interplay of various factors influencing its severity. Anticipating the onset of circulation loss is imperative to mitigate its impacts, minimise costs, and reduce risks to personnel and the environment.In this study, an innovative machine learning approach employing multigene genetic algorithms is utilised to analyse a dataset of 16,970 drilling datasets from 61 wells within the Marun oil field, located in Iran, where severe loss of circulation occurred. Geological characteristics, operational drilling parameters, and the properties of the drilling fluid were all considered. The dataset encompasses 19 parameters, of which seven are chosen as inputs for predicting lost circulation incidents. These inputs are then employed to construct a predictive model, employing an 85:15 training-to-test data ratio. To assess the model's performance, unseen datasets are utilised.The novelty of this study lies in the proposed model's consideration of a concise set of relevant input parameters, particularly real-time surface drilling parameters that are easily accessible for every well. The model attains a remarkable level of prediction accuracy for fluid loss, as indicated by various performance indices. The results indicate a mean absolute error of 1.33, a root mean square error of 2.58, and a coefficient of determination of 0.968. The suggested prediction model is optimised not only for data reduction but also for universal prediction and compatibility with other existing platforms. Moreover, it aids drilling engineers in implementing suitable mitigation strategies and designing optimal values for key operational surface parameters, both prior to and during drilling operations.

Logical genetic programming (LGP) application to water resources management.

Genetic programming (GP) is a variant of evolutionary algorithms (EA). EAs are general-purpose search algorithms. Yet, GP does not solve multi-conditional problems satisfactorily. This study improves the GP's predictive skill by development and integration of mathematical logical operators and functions to it. The proposed improvement is herein named logical genetic programming (LGP) whose performance is compared with that of GP using examples from the fields of mathematics and water resources. The results of the examples show the LGP's superior performance in both examples, with LGP producing improvements of 74 and 42% in the objective functions of the mathematical and water resources examples, respectively, when compared with the GP's results. The objective functions minimize the mean absolute error (MAE). The comparison of the LGP and GP results with alternative performance criteria demonstrate a better capability of the former algorithm in solving multi-conditional problems.

Rainstorm flash flood risk assessment using genetic programming: a case study of risk zoning in Beijing

In this study, a genetic programming (GP) algorithm is introduced to solve the symbolic regression problem for flash flood risk zoning in Beijing. GP operates in simulation of biological revolution and can avoid arbitrariness in risk estimates. Herein, this revolutionary computing searched for an appropriate model to best fit the training samples which comprise the data fields of the predictand of Ripley’s K-values to be the posterior risk, and the predictors of the rainstorm hazard index value (RHIV), physical vulnerability, terrain factor, impervious surface area, and population density. After generations of revolution, the optimal fit regressions for estimating the risk value were determined in the form of function (parse) trees. Also, the grid risk values were calculated using the deduced regression. The risk zoning map indicates that the risk values are higher in urban areas, which is reasonable in comparison with the distribution of historical flash flood events. With an explicit model structure, this symbolic regression manifests that the risk value is mainly determined by the RHIV and impervious land surface and is weakly correlated with the other risk factors, e.g., the physical vulnerability, the terrain factor, and population density. Our research demonstrates that GP in an artificial intelligence manner meets the needs of risk assessment in determining the optimal fit regressions and is a promising technique for future applications. Meanwhile, approaches are still available for improving the GP application in the risk assessment, e.g., considering the historical losses in posterior risk estimations and improvement in the sampling training data.

Prediction of inflows from dam catchment using genetic programming

Application of hydroinformatics tools for managing water resources is common in the water industry. Over the last few decades, several hydroinformatics tools including genetic programming (GP) have been developed and applied in hydrology. GP has been successfully applied for calibration of numerous event-based rainfall and runoff models. However, applying GP to predict long-term time series for the management of water resources is limited. This study demonstrates GP's application in long-term prediction of catchment runoff concerning a dam located in Oberon, New South Wales, Australia. The calibration showed excellent agreement between the observed and simulated flows recorded over 30 years. The model was then applied for the assessment of catchment yields for a future 100 years flows based on two assumed climatic scenarios.

Linear genetic programming application for successive-station monthly streamflow prediction

In recent decades, artificial intelligence (AI) techniques have been pronounced as a branch of computer science to model wide range of hydrological phenomena. A number of researches have been still comparing these techniques in order to find more effective approaches in terms of accuracy and applicability. In this study, we examined the ability of linear genetic programming (LGP) technique to model successive-station monthly streamflow process, as an applied alternative for streamflow prediction. A comparative efficiency study between LGP and three different artificial neural network algorithms, namely feed forward back propagation (FFBP), generalized regression neural networks (GRNN), and radial basis function (RBF), has also been presented in this study. For this aim, firstly, we put forward six different successive-station monthly streamflow prediction scenarios subjected to training by LGP and FFBP using the field data recorded at two gauging stations on Çoruh River, Turkey. Based on Nash–Sutcliffe and root mean squared error measures, we then compared the efficiency of these techniques and selected the best prediction scenario. Eventually, GRNN and RBF algorithms were utilized to restructure the selected scenario and to compare with corresponding FFBP and LGP. Our results indicated the promising role of LGP for successive-station monthly streamflow prediction providing more accurate results than those of all the ANN algorithms. We found an explicit LGP-based expression evolved by only the basic arithmetic functions as the best prediction model for the river, which uses the records of the both target and upstream stations.

TABU PROGRAMMING: A NEW PROBLEM SOLVER THROUGH ADAPTIVE MEMORY PROGRAMMING OVER TREE DATA STRUCTURES

Since the first appearance of the Genetic Programming (GP) algorithm, extensive theoretical and application studies on it have been conducted. Nowadays, the GP algorithm is considered one of the most important tools in Artificial Intelligence (AI). Nevertheless, several questions have been raised about the complexity of the GP algorithm and the disruption effect of the crossover and mutation operators. In this paper, the Tabu Programming (TP) algorithm is proposed to employ the search strategy of the classical Tabu Search algorithm with the tree data structure. Moreover, the TP algorithm exploits a set of local search procedures over a tree space in order to mitigate the drawbacks of the crossover and mutation operators. Extensive numerical experiments are performed to study the performance of the proposed algorithm for a set of benchmark problems. The results of those experiments show that the TP algorithm compares favorably to recent versions of the GP algorithm in terms of computational efforts and the rate of success. Finally, we present a comprehensive framework called Meta-Heuristics Programming (MHP) as general machine learning tools.

Genetic programming application to generate technical trading rules in stock markets

Technical trading rules can be generated from historical data for decision making in stock trading. In this study, genetic programming (GP) as an evolutionary algorithm has been applied to automatically generate such technical trading rules on individual stocks. In order to obtain more realistic trading rules, we have included transaction costs, dividends and splits in our GP model. Our model has been applied for nine Iranian companies listed on different activity sectors of Tehran Stock Exchange (TSE). Our results show that this model could generate profitable trading rules in comparison with buy and hold strategy for companies having frequent trading in the market. Also, the effect of the above mentioned parameters on trading rule's profitability are evaluated using three separate models.

Formulation of ELF magnetic fields’ effects on malondialdehyde level and myeloperoxidase activity in kidney using genetic programming

In vivo exposure effects of electromagnetic fields (EMFs) on various tissues of experiment animals have been investigated. In this sense, modeling and formulation of these biological effects have been of significant importance. In this study extremely low frequency (ELF) EMFs effects on malondialdehyde (MDA) level and myeloperoxidase (MPO) activity in kidney of guinea pigs exposed to 50 Hz magnetic fields of 1 mT, 2 mT and 3 mT have been presented. It has been planned to determine whether genetic programming (GP) is appropriate to analyze and formulate these biological effects. Consequently, it has been observed that GP can be effectively used to model MDA level and MPO activity. The performances of prediction of the proposed GP formulation versus actual experimental values are found to be quite satisfactory in terms of standard deviation and correlation coefficient. It is concluded that the GP application serves to form a database for the researchers in this field, without exposing tissues to EMF and without using too many guinea pigs.

Training Genetic Programming on Half a Million Patterns: An Example From Anomaly Detection

The hierarchical RSS-DSS algorithm is introduced for dynamically filtering large datasets based on the concepts of training pattern age and difficulty, while utilizing a data structure to facilitate the efficient use of memory hierarchies. Such a scheme provides the basis for training genetic programming (GP) on a data set of half a million patterns in 15 min. The method is generic, thus, not specific to a particular GP structure, computing platform, or application context. The method is demonstrated on the real-world KDD-99 intrusion detection data set, resulting in solutions competitive with those identified in the original KDD-99 competition, while only using a fraction of the original features. Parameters of the RSS-DSS algorithm are demonstrated to be effective over a wide range of values. An analysis of different cost functions indicates that hierarchical fitness functions provide the most effective solutions.

Forecasting with computer-evolved model specifications: a genetic programming application

This paper uses genetic programming (GP) to evolve model specifications of time series data. GP is a computerized random search optimization algorithm that assembles equations until it identifies the fittest one. The technique is applied here to artificially simulated data first then to real-world sunspot numbers. One-step-ahead forecasts produced by the fittest of computer-evolved models are evaluated and compared with alternatives. The results suggest that GP may produce reasonable forecasts if their user selects appropriate input variables and comprehends the process investigated. Further, the technique appears promising in forecasting noisy complex series perhaps better than other existing methods. It is suitable for decision makers who set high priority on obtaining accurate forecasts rather than on probing into and approximating the underlying data generating process. Scope and purpose This paper contains a brief introduction and an evaluation of the use of genetic programming (GP) in forecasting time series. GP is a computerized random search optimization technique based upon Darwin's theory of evolution. The algorithm is first applied to model and forecast artificially simulated linear and nonlinear time series. Results are used to evaluate the effectiveness of GP as a forecasting technique. It is then applied to model and forecast sunspot numbers—the most frequently analyzed and forecasted series. An autoregressive and a threshold nonlinear dynamical systems to capture the dynamics of the irregular sunspot numbers’ cycle were tested using GP. The latter delivered estimated equations yielding the lowest mean square error ever reported for the series. This paper demonstrates that GP's forecasting capabilities depend on the structure and complexity of the process to model. Skills and intuition of GP's user are its limitation.