genetic algorithm

Among the topics in the field of software engineering, energy efficiency is an influential factor in software development and maintenance, especially for battery-limited devices. Although software refactoring can improve software quality, recent studies suggest that applying some refactoring operators may lead to conflicts with energy consumption and execution time of Android applications. In this paper, we analyze the impact of code refactoring for eight Android/Java bad code smells and anti-patterns. To conduct the studies and obtain the results, we use a testbed of five real and one synthetic Android applications. In the first step, we measure energy consumption, execution time and quality design of application before and after refactoring. The results show that in some cases, refactoring leads to a decrease in energy consumption and execution time, and in others, it increases energy consumption and application execution time. We then propose a novel refactoring recommendation approach based on evolutionary multi-objective optimization that accounts for energy consumption, execution time and refactoring effort for Android/Java anti-patterns.  For this purpose, we use Nondominated Sorting Genetic Algorithm-II (NSGA-II) with three objectives 1) energy consumption, 2) execution time, and 3) refactoring effort. The obtained results show that this approach can generate refactoring recommendations with a median precision of 65% and 76% for improving energy and execution time, respectively, while the median of removed antipatterns in testbed applications is 42%.

The global warming and climate change due to the increase in greenhouse gas emissions are the vital and challenging issues in this period of human life. Human’s effort to address this problem has led to provide several solutions including net-zero carbon, i.e., carbon capture and storage and energy transition i.e., reducing or even replacing fossil fuels with renewable energy sources such as solar, wind, hydro etc. However, climate dependence and the highly fluctuating nature of clean energy production from these sources require a large-scale storage system to continuously meet energy demand. Hydrogen as an energy carrier, and underground hydrogen storage (UHS) hold potential for sustainable supply of a large amount of energy in the peak of energy consumption. Hydrogen has different characteristics and dynamical behaviors in the porous media compared to other gases. Thus it is required to create complex compositional models and perform time-consuming simulations to seek for the best gas storage scenario based on operational parameters including and cushion gas and working gas injection/production flowrates, perforations and cushion gas composition. In this study, the process of UHS in a depleted gas reservoir was simulated. Due to the reservoir remaining gas saturation and phase behavior of fluids in porous media, both hydrogen recovery factor and purity are considered as target variables. Next, the design of experiment methods (e.g., Latin hypercube) was utilized to generate the required train and test subsets for artificial neural network model. The feed-forward model with 10 neurons and sigmoid activation function as a smart proxy model with accuracy equal to 0.97 and 0.94, respectively, for training and testing subsets provided the best performance for predicting the hydrogen purity and recovery factor or the target parameters in the multi-objective optimization process of decision variables by genetic algorithm. The optimum solutions, i.e., Pareto front for the decision variables showed the dominant percentage of nitrogen gas with the base gas composition of 75, 20 and 5% respectively for nitrogen, carbon dioxide and methane. Moreover, production from top perforations and hydrogen injection in the lower part of the preformation were determined as optimal conditions. The implemented procedure in this paper can be used for UHS field studies and fast decision making in the large-scale energy storage operations.

In recent years, due to the optimal geometry and lower pressure drop, diffusion control vanes have significant applications, especially in the aviation industry and in subsonic and transsonic conditions. In the current research, the airfoil of the axial stator compressor section designed in the National Aerospace Laboratory of India has been selected as the basic geometry. The goal of optimization is to minimize the pressure drop of the entire fluid flow and consequently reduce the drop rate. The working method in this research is the change in the profile geometry of the blade by changing the parameters of the parsec method, which leads to the creation of new geometries at each stage of the code execution. The used optimization method is developed based on Genetic Algorithm. For the aerodynamic analysis of the generated geometry in each step and extracting the total pressure drop value, the MATLAB code is coupled with Ansys software and in each step, after numerical solution for each generated geometry, the total pressure drop value is extracted and returned to the code. Finally, the work output of the vane is more optimal and with a lower pressure drop, which is finally compared with the original vane and introduced as a suitable alternative. The total pressure drop between inlet and outlet in the optimized vane has decreased by 18% compared to the original vane, and the mass flow rate has also increased by 0.083 kg/s, which is a significant amount. The improvement of various aerodynamic characteristics such as Mach number distribution and pressure and drop coefficients can also be seen between the two basic and optimized blades, which is detailed at the end of the article

With the increasing use of Internet of Things in daily life and especially in industry, improving efficiency and delay time with the help of data offloading is one of the goals of these issues. Controlling these factors will improve energy consumption and longer use of things batteries. In this article, the method is introduced to improve sensor data processing and edge and cloud computing in industrial Internet of Things systems. The architecture is considered in accordance with the real world; in this architecture, edge servers with computing capabilities at the edge of the network, especially Used in base stations. Delay-sensitive requests can be forwarded to nearby edge servers through wireless channels, thereby reducing traffic in the core network and user data transmission latency, especially for data-intensive industrial applications. In the Industrial Internet of Things aims to manage network resources, transfer calculations and minimize energy consumption in Internet of Things devices by guaranteeing the freshness of sensor data. The network environment and input tasks are variable with time. In this article, the environment of the problem and its limitations are expressed with formulas. This problem has been solved using the proposed genetic algorithm and reinforcement learning. The proposed solution has improved the dynamic environment of the problem for offloading data and tasks by considering energy and transferring calculations and data by considering their freshness. The results show an average improvement of 40% compared to the previous methods.

One of the most significant aspects of train timetable optimization is determining the arrival dwell time in the subway system. A good timetable improves the efficiency of trains and resources, reducing passenger wait times. This study provides a timetable optimization model to reduce passenger journey time. In this model, train movement between two stations is divided into three phases: accelerating, coasting, and breaking, with random numbers used for running duration in each phase due to stochastic delays in crowded stations. First, we develop a stochastic integer programming model that includes headway and dwell time. The Scenario-Based Uncertain Programming Approach is then utilized to simplify. Finally, numerical examples are provided for Tehran's metro system. The model's optimal solution is determined via a genetic algorithm. The results demonstrate that the model can lower the passenger waiting time when compared to the current timetable.

The present study presents a three-tier model with incomplete and uncertain information of supply chain needs, benefits and services. Objectives of the issue include determining the best decision to determine the optimal order amount and shortage for the manufacturer and the selling price of each player according to the shortage, discount and maintenance costs, purchase and marketing to achieve maximum revenue, minimum costs and The sum is the maximum possible profit for all the players participating in the chain. To solve the model, Gamz software and meta-heuristic algorithms have been used and finally, Given the complexity of the complexity of closed-loop supply chain problems, the problem ahead cannot be solved in a reasonable time for real-world dimensions. Therefore, to solve it, the meta-heuristic approach in the form of genetic algorithms and optimization of particle aggregation and the combination of these two algorithms have been used. The results show that the combined algorithm of genetics and particle swarming has a better situation compared to genetic and particle swarming algorithms.

In this research, a deep reinforcement learning algorithm is proposed for the cellular manufacturing system problem considering the costs of delay and rejection of orders. Orders with different characteristics including revenue, lead time, delivery date, and delay cost are dynamically entered into the system at different times. Due to the limited capacity of the system, it is not possible to accept all orders and some of them must be rejected at the time of entry to enable timely execution of other orders. A mathematical model with two objectives of maximizing profit and minimizing the number of rejected orders is presented and a deep reinforcement learning algorithm is used to solve this problem. The proposed algorithm is compared with the algorithms available in the literature in different categories of example problems and real problems and its efficiency is proven. The results show a 36.3% advantage in profit and 13.87% in the number of accepted orders. Also, by accepting 1% more orders, the profit decreases by 2.7% on average

This study aims to develop a mathematical model for flexible job shop scheduling. The main focus is on optimizing three

the makespan, the maximum machine workload, and the total workload. The ultimate goal is to enhance productivity and flexibility in manufacturing systems.

Two metaheuristic algorithms, NSGA-II and MOGWO, were used to solve the model. The model was first validated on a small scale, and then sensitivity analysis was conducted on larger instances. The performance of the algorithms was compared based on accuracy and solution quality metrics.

Results indicated that MOGWO performs better in medium-sized problems, while in large-scale cases, the difference between the two algorithms is not significant. The highest sensitivity among the objectives was observed with respect to production and maintenance costs. Additionally, a resource allocation pattern and optimal sequence of operations were derived.

 The originality of this research lies in developing and applying a multi-objective mathematical model for flexible job shop scheduling, considering real-world constraints such as costs and resource limitations. The simultaneous use and detailed comparison of NSGA-II and MOGWO across different problem sizes is another contribution. Furthermore, the proposed operational pattern improves the applicability of the results in industrial environments.

The steel plate shear wall system is one of the most effective lateral load-resisting systems, offering significant behavioral advantages. However, existing design methods for this system are often overly conservative, leading to economic inefficiencies. This study aims to develop structural designs that ensure both effective safety and cost-efficiency. To achieve this, a genetic optimization algorithm is utilized. Initially, a numerical model is developed and validated using the OpenSees software. This model is then integrated with an optimization code implemented in MATLAB, which iteratively adjusts design parameters to determine optimal cross-sectional areas for 4-, 8-, and 12-story structures. These optimized designs are subsequently compared to conventionally designed models created through a trial-and-error approach. The findings demonstrate that optimization reduces structural weight, enhances performance, and mitigates the conservatism inherent in traditional design methods.

This research introduces the optimal design for underwater sound-absorbing linings. To address the time and computational cost of the optimization process, a combined approach of the genetic algorithm and a surrogate model using artificial neural networks was implemented. The results conclusively demonstrate that the best cavity shape for the low-frequency domain is conical with an apex near the surface of the entering wave. In contrast, a conical shape in the opposite direction of the low-frequency case is optimal for the entire frequency range. Selecting the best frequency range is crucial for the design and optimization process. The optimal response significantly outperformed other randomly established linings, validating the chosen approach. Additionally, increasing the number of optimization variables will undoubtedly alter the optimal solution, albeit with a considerable increase in problem complexity.