Rapid Prototype Development Approach for Genetic Programming

Genetic Programming (GP) is an important approach to deal with complex problem analysis and modeling, and has been applied in a wide range of areas. The development of GP involves various aspects, including design of genetic operators, evolutionary controls and implementations of heuristic strategy, evaluations and other mechanisms. When designing genetic operators, it is necessary to consider the possible limitations of encoding methods of individuals. And when selecting evolutionary control strategies, it is also necessary to balance search efficiency and diversity based on representation characteristics as well as the problem itself. More importantly, all of these matters, among others, have to be implemented through tedious coding work. Therefore, GP development is both complex and time-consuming. To overcome some of these difficulties that hinder the enhancement of GP development efficiency, we explore the feasibility of mutual assistance among GP variants, and then propose a rapid GP prototyping development method based on πGrammatical Evolution (πGE). It is demonstrated through regression analysis experiments that not only is this method beneficial for the GP developers to get rid of some tedious implementations, but also enables them to concentrate on the essence of the referred problem, such as individual representation, decoding means and evaluation. Additionally, it provides new insights into the roles of individual delineations in phenotypes and semantic research of individuals.

Efficient Graph-based Genetic Programming Representation with Multiple Outputs

In this work, we explore and study the implication of having more than one output on a genetic programming （GP） graph-representation. This approach, called multiple interactive outputs in a single tree （MIOST）, is based on two ideas. First, we defined an approach, called interactivity within an individual （IWI）, which is based on a graph-GP representation. Second, we add to the individuals created with the IWI approach multiple outputs in their structures and as a result of this, we have MIOST. As a first step, we analyze the effects of IWI by using only mutations and analyze its implications （i.e., presence of neutrality）. Then, we continue testing the effectiveness of IWI by allowing mutations and the standard GP crossover in the evolutionary process. Finally, we tested the effectiveness of MIOST by using mutations and crossover and conducted extensive empirical results on different evolvable problems of different complexity taken from the literature. The results reported in this paper indicate that the proposed approach has a better overall performance in terms of consistency reaching feasible solutions.

Application of numerical modeling and genetic programming to estimate rock mass modulus of deformation

Estimation of the rock mass modulus of deformation(Em)is one of the most important design parameters in designing many structures in and on rock.This parameter can be obtained by in situ tests,empirical relations between deformation modulus and rock mass classifcation,and estimating from laboratory tests results.In this paper,a back analysis calculation is performed to present an equation for estimation of the rock mass modulus of deformation using genetic programming(GP)and numerical modeling.A database of 40,960 datasets,including vertical stress(rz),horizontal to vertical stresses ratio(k),Poisson’s ratio(m),radius of circular tunnel(r)and wall displacement of circular tunnel on the horizontal diameter(d)for input parameters and modulus of deformation for output,was established.The selected parameters are easy to determine and rock mass modulus of deformation can be obtained from instrumentation data of any size circular galleries.The resulting RMSE of 0.86 and correlation coeffcient of97%of the proposed equation demonstrated the capability of the computer program(CP)generated by GP.

Improved genetic algorithm for nonlinear programming problems

An improved genetic algorithm（IGA） based on a novel selection strategy to handle nonlinear programming problems is proposed.Each individual in selection process is represented as a three-dimensional feature vector which is composed of objective function value,the degree of constraints violations and the number of constraints violations.It is easy to distinguish excellent individuals from general individuals by using an individuals＇ feature vector.Additionally,a local search（LS） process is incorporated into selection operation so as to find feasible solutions located in the neighboring areas of some infeasible solutions.The combination of IGA and LS should offer the advantage of both the quality of solutions and diversity of solutions.Experimental results over a set of benchmark problems demonstrate that IGA has better performance than other algorithms.

Solution for integer linear bilevel programming problems using orthogonal genetic algorithm

An integer linear bilevel programming problem is firstly transformed into a binary linear bilevel programming problem, and then converted into a single-level binary implicit programming. An orthogonal genetic algorithm is developed for solving the binary linear implicit programming problem based on the orthogonal design. The orthogonal design with the factor analysis, an experimental design method is applied to the genetic algorithm to make the algorithm more robust, statistical y sound and quickly convergent. A crossover operator formed by the orthogonal array and the factor analysis is presented. First, this crossover operator can generate a smal but representative sample of points as offspring. After al of the better genes of these offspring are selected, a best combination among these offspring is then generated. The simulation results show the effectiveness of the proposed algorithm.

Exponential distribution-based genetic algorithm for solving mixed-integer bilevel programming problems

Two classes of mixed-integer nonlinear bilevel programming problems are discussed. One is that the follower＇s functions are separable with respect to the follower＇s variables, and the other is that the follower＇s functions are convex if the follower＇s variables are not restricted to integers. A genetic algorithm based on an exponential distribution is proposed for the aforementioned problems. First, for each fixed leader＇s variable x, it is proved that the optimal solution y of the follower＇s mixed-integer programming can be obtained by solving associated relaxed problems, and according to the convexity of the functions involved, a simplified branch and bound approach is given to solve the follower＇s programming for the second class of problems. Furthermore, based on an exponential distribution with a parameter λ, a new crossover operator is designed in which the best individuals are used to generate better offspring of crossover. The simulation results illustrate that the proposed algorithm is efficient and robust.

Genetic Programming Approach for Predicting Surface Subsidence Induced by Mining

The surface subsidence induced by mining is a complex problem, which is related with many complex and uncertain factors. Genetic programming （GP） has a good ability to deal with complex and nonlinear problems, therefore genetic programming approach is propesed to predict mining induced surface subsidence in this article. First genetic programming technique is introduced, second, surface subsidence genetic programming model is set up by selecting its main affective factors and training relating to practical engineering data, and finally, predictions are made by the testing of data, whose results show that the relative error is approximately less than 10%, which can meet the engineering needs, and therefore, this proposed approach is valid and applicable in predicting mining induced surface subsidence. The model offers a novel method to predict surface subsidence in mining.

Orthogonal genetic algorithm for solving quadratic bilevel programming problems

A quadratic bilevel programming problem is transformed into a single level complementarity slackness problem by applying Karush-Kuhn-Tucker（KKT） conditions.To cope with the complementarity constraints,a binary encoding scheme is adopted for KKT multipliers,and then the complementarity slackness problem is simplified to successive quadratic programming problems,which can be solved by many algorithms available.Based on 0-1 binary encoding,an orthogonal genetic algorithm,in which the orthogonal experimental design with both two-level orthogonal array and factor analysis is used as crossover operator,is proposed.Numerical experiments on 10 benchmark examples show that the orthogonal genetic algorithm can find global optimal solutions of quadratic bilevel programming problems with high accuracy in a small number of iterations.

Use Genetic Programming to Rank Web Images

Web image retrieval is a challenging task. One central problem of web image retrieval is to rank a set of images according to how well they meet the user information need. The problem of learning to rank has inspired numerous approaches to resolve it in the text information retrieval, related work for web image retrieval, however, are still limited. We focus on the problem of learning to rank images for web image retrieval, and propose a novel ranking model, which employs a genetic programming architecture to automatically generate an effective ranking function, by combining various types of evidences in web image retrieval, including text information, image visual content features, link structure analysis and temporal information. The experimental results show that the proposed algorithms are capable of learning effective ranking functions for web image retrieval. Significant improvement in relevancy obtained, in comparison to some other well-known ranking techniques, in terms of MAP, NDCG@n and D@n.

Genetic programming for predictions of effectiveness of rolling dynamic compaction with dynamic cone penetrometer test results

Rolling dynamic compaction (RDC),which employs non-circular module towed behind a tractor,is an innovative soil compaction method that has proven to be successful in many ground improvement applications.RDC involves repeatedly delivering high-energy impact blows onto the ground surface,which improves soil density and thus soil strength and stiffness.However,there exists a lack of methods to predict the effectiveness of RDC in different ground conditions,which has become a major obstacle to its adoption.For this,in this context,a prediction model is developed based on linear genetic programming (LGP),which is one of the common approaches in application of artificial intelligence for nonlinear forecasting.The model is based on in situ density-related data in terms of dynamic cone penetrometer (DCP) results obtained from several projects that have employed the 4-sided,8-t impact roller (BH-1300).It is shown that the model is accurate and reliable over a range of soil types.Furthermore,a series of parametric studies confirms its robustness in generalizing data.In addition,the results of the comparative study indicate that the optimal LGP model has a better predictive performance than the existing artificial neural network (ANN) model developed earlier by the authors.

Prediction of Concrete Faced Rock Fill Dams Settlements Using Genetic Programming Algorithm

In the present study a Genetic Programing model (GP) proposed for the prediction of relative crest settlement of concrete faced rock fill dams. To this end information of 30 large dams constructed in seven countries across the world is gathered with their reported settlements. The results showed that the GP model is able to estimate the dam settlement properly based on four properties, void ratio of dam’s body (e), height (H), vertical deformation modulus (Ev) and shape factor (Sc) of the dam. For verification of the model applicability, obtained results compared with other research methods such as Clements’s formula and the finite element model. The comparison showed that in all cases the GP model led to be more accurate than those of performed in literature. Also a proper compatibility between the GP model and the finite element model was perceived.

Genetic Algorithm for Solving Quadratic Bilevel Programming Problem

By applying Kuhn-Tucker condition the quadratic bilevel programming, a class of bilevel programming, is transformed into a single level programming problem, which can be simplified by some rule. So we can search the optimal solution in the feasible region, hence reduce greatly the searching space. Numerical experiments on several literature problems show that the new algorithm is both feasible and effective in practice.

Inference of General Mass Action-Based State Equations for Oscillatory Biochemical Reaction Systems Using <i>k</i>-Step Genetic Programming

Systems biology requires the development of algorithms that use omics data to infer interaction networks among biomolecules working within an organism. One major type of evolutionary algorithm, genetic programming (GP), is useful for its high heuristic ability as a search method for obtaining suitable solutions expressed as tree structures. However, because GP determines the values of parameters such as coefficients by random values, it is difficult to apply in the inference of state equations that describe oscillatory biochemical reaction systems with high nonlinearity. Accordingly, in this study, we propose a new GP procedure called “k-step GP” intended for inferring the state equations of oscillatory biochemical reaction systems. The k-step GP procedure consists of two algorithms: 1) Parameter optimization using the modified Powell method—after genetic operations such as crossover and mutation, the values of parameters such as coefficients are optimized by applying the modified Powell method with secondary convergence. 2) GP using divided learning data—to improve the inference efficiency, imposes perturbations through the addition of learning data at various intervals and adaptations to these changes result in state equations with higher fitness. We are confident that k-step GP is an algorithm that is particularly well suited to inferring state equations for oscillatory biochemical reaction systems and contributes to solving inverse problems in systems biology.

A predictive equation for residual strength using a hybrid of subset selection of maximum dissimilarity method with Pareto optimal multi-gene genetic programming

More accurate and reliable estimation of residual strength friction angle(/r)of clay is crucial in many geotechnical engineering applications,including riverbank stability analysis,design,and assessment of earthen dam slope stabilities.However,a general predictive equation for/r,with applicability in a wide range of effective parameters,remains an important research gap.The goal of this study is to develop a more accurate equation for/r using the Pareto Optimal Multi-gene Genetic Programming(POMGGP)approach by evaluating a comprehensive dataset of 290 experiments compiled from published literature databases worldwide.A new framework for integrated equation derivation proposed that hybridizes the Subset Selection of Maximum Dissimilarity Method(SSMD)with Multi-gene Genetic Programming(MGP)and Pareto-optimality(PO)to find an accurate equation for/r with wide range applicability.The final predictive equation resulted from POMGGP modeling was assessed in comparison with some previously published machine learning-based equations using statistical error analysis criteria,Taylor diagram,revised discrepancy ratio(RDR),and scatter plots.Base on the results,the POMGGP has the lowest uncertainty with U95=2.25,when compared with Artificial Neural Network(ANN)(U95=2.3),Bayesian Regularization Neural Network(BRNN)(U95=2.94),Levenberg-Marquardt Neural Network(LMNN)(U95=3.3),and Differential Evolution Neural Network(DENN)(U95=2.37).The more reliable results in estimation of/r derived by POMGGP with reliability 59.3%,and resiliency 60%in comparison with ANN(reliability=30.23%,resiliency=28.33%),BRNN(reliability=10.47%,resiliency=10.39%),LMNN(reliability=19.77%,resiliency=20.29%)and DENN(reliability=27.91%,resiliency=24.19%).Besides the simplicity and ease of application of the new POMGGP equation to a broad range of conditions,using the uncertainty,reliability,and resilience analysis confirmed that the derived equation for/r significantly outperformed other existing machine learning methods,including the ANN,BRNN,LMNN,and DENN equations。

Defocus Blur Segmentation Using Genetic Programming and Adaptive Threshold

Detection and classification of the blurred and the non-blurred regions in images is a challenging task due to the limited available information about blur type,scenarios and level of blurriness.In this paper,we propose an effective method for blur detection and segmentation based on transfer learning concept.The proposed method consists of two separate steps.In the first step,genetic programming(GP)model is developed that quantify the amount of blur for each pixel in the image.The GP model method uses the multiresolution features of the image and it provides an improved blur map.In the second phase,the blur map is segmented into blurred and non-blurred regions by using an adaptive threshold.A model based on support vector machine(SVM)is developed to compute adaptive threshold for the input blur map.The performance of the proposed method is evaluated using two different datasets and compared with various state-of-the-art methods.The comparative analysis reveals that the proposed method performs better against the state-of-the-art techniques.

Estimating uniaxial compressive strength of rocks using genetic expression programming

The aim of this paper is to estimate the uniaxial compressive strength（UCS） of rocks with different characteristics by using genetic expression programming（GEP）.For this purpose,five different types of rocks including basalt and ignimbrite（black,yellow,gray,brown） were prepared.Values of unit weight,water absorption by weight,effective porosity and UCS of rocks were determined experimentally.By using these experimental data,five different GEP models were developed for estimating the values of UCS for different rock types.Good agreement between experimental data and predicted results is obtained.

A High Precision Comprehensive Evaluation Method for Flood Disaster Loss Based on Improved Genetic Programming

Precise comprehensive evaluation of flood disaster loss is significant for the prevention and mitigation of flood disasters. Here, one of the difficulties involved is how to establish a model capable of describing the complex relation between the input and output data of the system of flood disaster loss. Genetic programming (GP) solves problems by using ideas from genetic algorithm and generates computer programs automatically. In this study a new method named the evaluation of the grade of flood disaster loss (EGFD) on the basis of improved genetic programming (IGP) is presented (IGP-EGFD). The flood disaster area and the direct economic loss are taken as the evaluation indexes of flood disaster loss. Obviously that the larger the evaluation index value, the larger the corresponding value of the grade of flood disaster loss is. Consequently the IGP code is designed to make the value of the grade of flood disaster be an increasing function of the index value. The result of the application of the IGP-EGFD model to Henan Province shows that a good function expression can be obtained within a bigger searched function space; and the model is of high precision and considerable practical significance. Thus, IGP-EGFD can be widely used in automatic modeling and other evaluation systems.

Hierarchical On-line Scheduling of Multiproduct Batch Plants with a Combined Approach of Mathematical Programming and Genetic Algorithm

In this contribution we present an online scheduling algorithm for a real world multiproduct batch plant. The overall mixed integer nonlinear programming (MINLP) problem is hierarchically structured into a mixed integer linear programming (MILP) problem first and then a reduced dimensional MINLP problem, which are optimized by mathematical programming (MP) and genetic algorithm (GA) respectively. The basis idea relies on combining MP with GA to exploit their complementary capacity. The key features of the hierarchical model are explained and illustrated with some real world cases from the multiproduct batch plants.

Comparison between dynamic programming and genetic algorithm for hydro unit economic load dispatch

The hydro unit economic load dispatch （ELD） is of great importance in energy conservation and emission reduction. Dynamic programming （DP） and genetic algorithm （GA） are two representative algorithms for solving ELD problems. The goal of this study was to examine the performance of DP and GA while they were applied to ELD. We established numerical experiments to conduct performance comparisons between DP and GA with two given schemes. The schemes included comparing the CPU time of the algorithms when they had the same solution quality, and comparing the solution quality when they had the same CPU time. The numerical experiments were applied to the Three Gorges Reservoir in China, which is equipped with 26 hydro generation units. We found the relation between the performance of algorithms and the number of units through experiments. Results show that GA is adept at searching for optimal solutions in low-dimensional cases. In some cases, such as with a number of units of less than 10, GA＇s performance is superior to that of a coarse-grid DP. However, GA loses its superiority in high-dimensional cases. DP is powerful in obtaining stable and high-quality solutions. Its performance can be maintained even while searching over a large solution space. Nevertheless, due to its exhaustive enumerating nature, it costs excess time in low-dimensional cases.

An adaptive genetic algorithm for solving bilevel linear programming problem

Bilevel linear programming, which consists of the objective functions of the upper level and lower level, is a useful tool for modeling decentralized decision problems. Various methods are proposed for solving this problem. Of all the algorithms, the ge- netic algorithm is an alternative to conventional approaches to find the solution of the bilevel linear programming. In this paper, we describe an adaptive genetic algorithm for solving the bilevel linear programming problem to overcome the difficulty of determining the probabilities of crossover and mutation. In addition, some techniques are adopted not only to deal with the difficulty that most of the chromosomes maybe infeasible in solving constrained optimization problem with genetic algorithm but also to improve the efficiency of the algorithm. The performance of this proposed algorithm is illustrated by the examples from references.