The Evolution of Genetic Algorithms: Towards Massive Parallelism

One of the important problems of sustainable logistics is routing vehicles in a sustainable manner, the green vehicle routing problem, or vehicle routing problems which aim to reduce CO2 emissions. In the literature research, it was seen that these problems were solved with heuristic, metaheuristic, or hyper-heuristic methods and hybrid approaches since they are in the NP-hard class. This work presents a parallel multi-process genetic algorithm that incorporates problem-specific genetic operators to minimize CO2 emissions in the capacity-constrained vehicle routing problem. Unlike previous research, the algorithm combines parallel computing with tailored genetic operators in order to enhance the diversity of solutions and speed up convergence. Genetic algorithm models were developed to minimize total distance, CO2 emissions, and both objectives simultaneously. Two genetic algorithm models were developed to minimize total distance and CO2 emissions. Experimental results using the reference CVRP examples such as A-n32-k5 and B-n44-k7 show that the proposed approach reduces CO2 emissions by 1.2% more than hybrid artificial bee colony optimization, 1.3% more than ant colony optimization, and 4% more than the traditional genetic algorithm. Experimental results using benchmark CVRP instances demonstrate that the proposed approach outperforms hybrid artificial bee colony optimization, ant colony optimization, and traditional genetic algorithms for most of the test cases. This is done by exploiting multi-core processors, and the parallel architecture has improved computational efficiency; the modules compare and update solutions against the global optimum. Results obtained show that prioritizing CO2 emissions as the only objective yields better results compared to multi-objective models. This study makes two significant contributions to the literature: (1) it introduces a novel parallel genetic algorithm framework optimized for CO2 emission reduction, and (2) it provides empirical evidence underscoring the advantages of emission-focused optimization in CVRP.

The main problem on community detection using traditional genetic algorithms (GA) lies in the slow speed of convergence. This paper attempts to apply parallel genetic algorithms (PGA) to explore community structure in complex networks in order to improve the efficiency of traditional genetic algorithms. Several different designing ways of PGA are discussed and compared. Experimental results based on the GN benchmark networks, LFR benchmark networks, and eight real-world networks, confirm the PGA with coarse-grained-master-slave hybrid model spends less time yet achieves higher accuracy than traditional genetic algorithms.

Summary form only given. This article gives a brief overview of theoretical advances, computing trends, applications and future perspectives in parallel genetic algorithms. It explains basic terms and behavior of (parallel) genetic algorithms. Genetic algorithms are easily parallelized algorithms, therefore two kinds of possible parallelism, data parallelism and control parallelism, are mentioned and described towards them. Parallelism of genetic algorithms brings many advantages and gains. Classifications of these algorithms are often based on the type of computing model, a walk strategy and the used computing machinery. Afterwards significant milestones in the theory with latest advances are briefly mentioned. Then current trends in parallel computing with stress computer architectures of parallel systems, interconnection topologies, operating systems, parallel (genetic) libraries and programming paradigms are reviewed shortly. The sufficient space is devoted to the latest applications of parallel genetic algorithms. After the discussion section, perspectives of the algorithms are predicted till the year 2005. The information in the article is segregated into two periods before and after the year 2000 in all chapters. The second period is more interesting and of higher importance, because it highlights recent research efforts and gives some hints about possible future trends. That is why we devote much space to the second period. As there is no such an overview of the recent period of parallel genetic algorithms, our investigation could be appealing and useful in many aspects.

In the genetic algorithm for compute-intensive task, the execution time of fitness function increases rapidly. As a result, when the population size or evolution population increases, the convergence speed of the algorithm is very slow. Based on the supercomputer platform, this paper designs and implements a parallel genetic algorithm architecture with dynamic migration strategy. A hybrid parallel genetic algorithm based on "coarse-grained and master-slave" two-level parallel architecture is proposed. The proposed architecture is composed of hybrid parallel programming models. Compared with the traditional genetic algorithm implemented on the simple MPI architecture, the proposed structure achieves better performance. Experiments show that the convergence speed of the algorithm is significantly improved.

This paper can be asserted that the use of parallel genetic algorithm can not only effectively improve the calculation speed and optimize the quality, but also can improve a lot of advantages. Reliability optimization for computer net works, subject s to cost constraints, is a NP-hard combinational problem. Reg ar ding a known network topology, the problem of choosing links and switchers among alternatives different in reliability and cost is settled by a Coarse-grained parallel genetic algorithm, which maximize the network availablity within a fixed budget. T he simulations on a dedicated cluster demonst rate that contracting to the sequential counterpart, o ur par allel GA improves the quality of plans greatly with an evident speed-up.

To overcome the disadvantages of traditional genetic algorithms, which easily fall to local optima, this paper proposes a hybrid genetic algorithm based on information entropy and game theory. First, a calculation of the species diversity of the initial population is conducted according to the information entropy by combining parallel genetic algorithms, including using the standard genetic algorithm (SGA), partial genetic algorithm (PGA) and syncretic hybrid genetic algorithm based on both SGA and PGA for evolutionary operations. Furthermore, with parallel nodes, complete-information game operations are implemented to achieve an optimum for the entire population based on the values of both the information entropy and the fitness of each subgroup population. Additionally, the Rosenbrock, Rastrigin and Schaffer functions are introduced to analyse the performance of different algorithms. The results show that compared with traditional genetic algorithms, the proposed algorithm performs better, with higher optimization ability, solution accuracy, and stability and a superior convergence rate.

This paper presents a crowding niche cellular genetic algorithm (referred to NCGA) aiming at solving the problems of local convergence and non-uniform population distribution in traditional genetic algorithm. The selecting operation in traditional genetic algorithm is improved by bringing in the concept of neighbors of cellular genetic algorithm, and the population distribution is greatly enhanced by introducing crowding niche mechanism, which betters the ability of global searching and helps to avoid the population local convergence. Meanwhile, the paper describes the crowding niche cellular genetic algorithm in details and compares it with simple genetic algorithm （SGA） and simple niche genetic algorithm （NGA）; the comparison results reveal that, NCGA outperforms the other two algorithms in terms of convergence rate and population diversity.

Aimed for the problems in traditional genetic algorithm of low search efficiency and easily falling into the local optimal solution, an improved genetic algorithm is proposed in this paper. So as to save the storage space, the simple one-dimensional code method is adopted to replace the method of complex two-dimensional coding. In the design of genetic operators, many operations such as crossover and mutation are redefined to avoid getting into the local optimum. Then the two fitness functions of collision-free path and the shortest distance are fused into one for the following genetic optimization. In the case of the same population parameters, 100 trials are respectively developed with the method of improved genetic algorithm and traditional genetic algorithm. Among them, the improved genetic algorithm to search the optimal path gets to 95 times, and the shortest path is 20.9706. Besides, the average searching time takes up 217ms. While the number of traditional method to search for the optimal path reaches up to 62 times, the shortest path can be 25.0711, and the average searching time needs 345ms. So compared to the tests results referred above, the improved genetic algorithm is more efficient and can get a better solution than the traditional genetic algorithm.

A new genetic algorithm (GA), called intelligent GA, is proposed by the inclusion of a fractional factorial design in the crossover operator to determine the best combination of design variables. That is, in the proposed GA, the chromosomes of children are generated via an intelligent crossover process with factorial experiments after completing the selection of GA operators. This gene mating process differs from the traditional GA, where the child chromosome is generally generated via randomization. Accordingly, the traditional GA incorporating the technique of fractional factorial design into the crossover operator markedly enhances the performance in terms of evolutionary efficiency. Another GA, called Taguchi-GA, that incorporates the Taguchi orthogonal arrays into the crossover operator is also developed to compare the performance of the presented intelligent GA. Four representative test cases, including two multimodal functions, a nonlinear dynamic control function, and a profile fitting of a high lift airfoil are performed via the traditional GA, micro-GA, Taguchi-GA, and present GA to assess the capacity and efficiency of the proposed GA. The computational solutions of the convergence history and optimal objective function for the test cases demonstrate that the proposed intelligent GA outperforms the other three GAs. Moreover, the presented new intelligent GA and the other three GAs are applied to the aerodynamic optimization of two airplanes, namely, an advanced defense airplane and an F-16A fighter. The redesigned wing planform is obtained by computations made with an aerodynamic flow solver, and the present GA depicts the optimal swept angle, aspect ratio, and taper ratio. The drag distribution and convergence history show that the proposed intelligent GA has a high evolutionary efficiency when applied to practical engineering problems.

Modern group communication based applications require multiple parameters to be considered for routing in a Cellular network. Traditional algorithms fail in the situations where these parameters frequently change due to the dynamism prevailing in the network. A new technique for topology discovery in these types of networks using ant colony optimization (ACO) has been proposed based on the restricted flooding principle. To provide a better quality of service in routing with multiple constraints, a genetic algorithm based routing has been proposed to find optimal routes within a shorter span of time than the traditional deterministic routing algorithms. Moreover, with the exponential growth in the number of mobile users, to enable a large number of users to participate in a group communication, a parallel genetic algorithm (GA) is proposed in this paper. Our simulation results show that the topology discovery using ant colony optimization is faster. The Call service rate using parallel genetic algorithm is more than that of sequential genetic algorithm and the Call blocking rate of parallel genetic algorithm is less than that of sequential genetic algorithm, for large number of routers in the network.

Genetic algorithms(GA), which are based on the idea of natural selection and survival of the fittest, have proven successful in solving complex problems that are not easily solved through conventional methods. Parallel genetic algorithm(PGA) is an extension of the classical GA. The important aspect in PGA is migration and GA operation. This paper presents PGAs that use fuzzy algorithm and natural select concept to migration of subpopulations. Fuzzy algorithm is used to change GA operators. Experimental results show that the proposed methods exhibit good performance compared to the classical method.

The process of Medical image registration is essentially to find a optimal transformation between two medical images, Because of the advantages and disadvantages of various optimization algorithm is relative, this paper proposed by the simplex algorithm and genetic algorithm hybrid optimization algorithm combining for registration parameter optimization, analysed and compared with the traditional genetic algorithm through experiment, verify the simplex genetic algorithm whether in registration precision or in registration efficiency are obviously better than the traditional genetic algorithm, Is a kind of practical optimization algorithm.

Floorplanning is an important problem in very large-scale integrated-circuit (VLSI) design automation as it determines the performance, size, yield and reliability of VLSI chips. From the computational point of view, floorplan area minimization is an NP-hard problem. This paper presents a parallel genetic algorithm (GA) for floorplan area optimization. The parallel GA is based an island model with an asynchronous migration mechanism, and is implemented using Web services and multithreading technologies. The parallel GA is compared with a sequential GA that the parallel GA is based on. Experimental results show that the parallel GA can produce better results than the sequential GA when they use the same amount of computing resources. In addition, since the number of islands and migration interval are two important parameters that directly affect the performance of island-based parallel GAs, the impact of the two parameters on the performance of the parallel GA are empirically studied in this paper.

Genetic algorithms offer very good performances for solving large optimization problems, especially in the domain of error-correcting codes. However, they have a major drawback related to the time complexity and memory occupation when running on a uniprocessor computer. This paper proposes a parallel decoder for linear block codes, using parallel genetic algorithms (PGA). The good performance and time complexity are confirmed by theoretical study and by simulations on BCH(63,30,14) codes over both AWGN and flat Rayleigh fading channels. The simulation results show that the coding gain between parallel and single genetic algorithm is about 0.7 dB at BER = 10﹣5 with only 4 processors.

This paper proposes a Social Genetic Algorithm (SGA) that includes a transformation function that has ability to improve search efficiency. The SGA is different from the Traditional Genetic Algorithm (TGA) approaches, as it allows refinement of the TGA parameters for the selections of operators in each generation with two functions: optimization of crossover rate and optimization of mutation rate. In this paper, a new function that optimizes gene relationship has been introduced to advance the evolution capability and flexibility of SGA in searching complex and large solution space. Our proposed approach has been evaluated using simulation models. The simulation results have shown that SGA outperforms TGA in improving search efficiency. The contribution of the proposed approach is a dynamic and adaptive methodology, which has ability to improve efficiency.