Non-dominated Sorting Genetic Algorithm II Research Articles

Research on Multi-Objective Optimization Methods of Urban Rail Train Automatic Driving Based on NSGA-II

In order to improve the control performance of automatic train operation (ATO) in urban rail trains, five typical operating sequences of urban rail trains were studied. Under the condition of meeting the safety and comfort principles of train operation, a train dynamics model was established to achieve the goals of low energy consumption, short running time, and high stopping accuracy in urban rail transit trains. In the process of finding a multi-objective solution to this problem, the Non-dominated Sorting Genetic Algorithm II (NSGA-II) was used with an elite retention strategy, and the optimal Pareto multi-objective solution set was sought. In the process of optimal solution weight assignment, the hierarchical analysis Mahalanobis distance method, which combines subjective and objective analysis, was used. Finally, taking the Beijing Yizhuang subway line as the background design example, the simulation verified the effectiveness and feasibility of the algorithm and obtained high-quality automatic train driving curves under various working conditions. This research has important reference significance for the actual operation of automatic driving in urban rail trains.

Multi‐criteria hydraulic turbine optimization using a genetic algorithm and trust‐region postprocessing

AbstractA two‐stage, multi‐objective optimization approach is proposed to improve the overall efficiency of the time‐consuming and computationally expensive optimization process for hydraulic turbines. The purpose is to optimize the shape of the turbine blades, designed using 30 parameters, in order to maximize its efficiency and minimize both the cavitation volume and the deviation between the desired and the calculated design head. In the first step of the routine, the genetic algorithm “Non‐Dominated Sorting Genetic Algorithm II (NSGA‐II)” is used to generate an initial approximation of the Pareto front that covers a wide variety of possible solutions, that is, optimal trade‐offs between the conflicting objectives. Subsequently, to ensure convergence, the trust region optimizer “Constrained Multiobjective Problem Optimizer with Model Information to Save Evaluations (CoMPrOMISE)” is used, which is initialized with individuals selected from the Pareto set generated by the genetic algorithm.

Optimization of Flange Design for Engine Assembly Stand using RSM and NSGA-II Based on FEA Data

This article presents the results of a research study that focused on optimizing the flange design for engine assembly stands using the Response Surface Methodology (RSM) and Nondominated Sorting Genetic Algorithm II (NSGA-II) based on finite element analysis (FEA) simulation data. The study investigated the impact of three main dimensions (d1, d2, d3) of the flange as independent variables on the mass (m, kg), Von Mises stress (V, MPa), and displacement (D, mm) as dependent variables. The regression models developed using the RSM exhibited high R2 values of 0.9896, 0.998, and 0.9997 for m, V, and D, respectively. The multi-objective optimization results obtained through NSGA-II yielded 39 Pareto solutions, with d1 ranging from 10 to 27.43 mm, d2 ranging from 30 to 50 mm, and d3 ranging from 75 to 85 mm. These values corresponded to m values ranging from 2.93 to 7.58 kg, V values ranging from 54.8 to 342.6 MPa, and D values ranging from 0.006 to 0.270 mm. For verification purposes, Solution No.17 was selected. The results showed that the redesigned flange's mass, Von Mises stress, and displacement deviated by 0.95%, 2.28%, and 1.14%, respectively, from the optimal values obtained through the optimization process. These findings provide strong evidence for the high reliability of the optimization method employed in this study.

Optimization of household medical waste recycling logistics routes: Considering contamination risks.

The escalating generation of household medical waste, a byproduct of industrialization and global population growth, has rendered its transportation and logistics management a critical societal concern. This study delves into the optimization of routes for vehicles within the household medical waste logistics network, a response to the imperative of managing this waste effectively. The potential for environmental and public health hazards due to improper waste disposal is acknowledged, prompting the incorporation of contamination risk, influenced by transport duration, waste volume, and wind velocity, into the analysis. To enhance the realism of the simulation, traffic congestion is integrated into the vehicle speed function, reflecting the urban roads' variability. Subsequently, a Bi-objective mixed-integer programming model is formulated to concurrently minimize total operational costs and environmental pollution risks. The complexity inherent in the optimization problem has motivated the development of the Adaptive Hybrid Artificial Fish Swarming Algorithm with Non-Dominated Sorting (AH-NSAFSA). This algorithm employs a sophisticated approach, amalgamating congestion distance and individual ranking to discern optimal solutions from the population. It incorporates a decay function to facilitate an adaptive iterative process, enhancing the algorithm's convergence properties. Furthermore, it leverages the concept of crossover-induced elimination to preserve the genetic diversity and overall robustness of the solution set. The empirical evaluation of AH-NSAFSA is conducted using a test set derived from the Solomon dataset, demonstrating the algorithm's capability to generate feasible non-dominated solutions for household medical waste recycling path planning. Comparative analysis with the Non-dominated Sorted Artificial Fish Swarm Algorithm (NSAFSA) and Non-dominated Sorted Genetic Algorithm II (NSGA-II) across metrics such as MID, SM, NOS, and CT reveals that AH-NSAFSA excels in MID, SM, and NOS, and surpasses NSAFSA in CT, albeit slightly underperforming relative to NSGA-II. The study's holistic approach to waste recycling route planning, which integrates cost-effectiveness with pollution risk and traffic congestion considerations, offers substantial support for enterprises in formulating sustainable green development strategies. AH-NSAFSA offers an eco-efficient, holistic approach to medical waste recycling, advancing sustainable management practices.

Performance analysis and multi-objective optimization for an integrated air separation, power generation, refrigeration and ice thermal storage system based on the LNG cold energy utilization

Against the backdrop of escalating resource depletion and the urgent quest for alternative sources, liquefied natural gas (LNG) is increasingly gaining prominence as a sustainable solution, particularly in refrigeration applications. However, its underutilization results in wasted resources. To efficiently harness the released cold energy from LNG gasification, this study proposes an integrated system comprising air separation, power generation, refrigeration, and ice thermal storage. The system undergoes optimization using the non-dominated sorting genetic algorithm II (NSGA-II) to determine the optimal operating parameters. The optimized system is comprehensively analyzed from energy, exergy, economic, and environmental perspectives. Results show that the system, with a 70t/h LNG capacity, achieves an energy efficiency of 42.52% and an exergy efficiency of 48.09%. Economically, the system incurs a cost of approximately 0.0711 $/kWh and can mitigate over 1.4504*107kg of CO2 emissions. Compared to traditional LNG utilization systems, the integrated system demonstrates a 22.32% improvement in energy efficiency, a 7.69% enhancement in exergy efficiency, and a cost reduction of 0.0049 $/kWh. In summary, this technologically advanced and economically viable system offers a significant alternative to optimize LNG cold energy utilization.

Enhancing International Logistics and Supply Chain Management: Deep Learning Strategies for Enhanced Route Planning and Warehouse Optimization

Logistics and Supply Chain Management (SCM) are both key areas in modern industry and commerce which need better route planning and warehouse optimization. Traditional methods that are in practice have often fall short in of addressing the dynamic complexities of modern logistics which results in inefficient travel times, fuel consumption, and space utilization. To counter these limitations this study introduces an integrated model that combines Long Short-Term Memory (LSTM) networks, Radial Basis Function Neural Networks (RBNN), and the Non-dominated Sorting Genetic Algorithm II (NSGA-II) for route planning and warehouse optimization. The proposed model employ LSTM to predict traffic patterns and RBNN to optimize space utilization in warehouse. The NSGA-II model is then utilized for multi-objective optimization of minimizing travel times and maximizing warehouse space utilization. In experiment analysis the proposed model achieved the highest accuracy and least variability in predictions, with mean MAE, RMSE, and MAPE values of 0.57, 1.12, and 5.9%, respectively.

Electric vehicle power system in intelligent manufacturing based on soft computing optimization

Soft computing technology has attracted extensive attention in computer engineering and automatic control domains because it can deal with uncertainties, fuzziness, and complex practical problems. This study adopts a Genetic Algorithm (GA) in soft computing technology to realize the cooperative optimization of electric vehicle's dynamic and economic performance. The advantage of soft computing technology lies in its adaptability to uncertainty, fuzziness, and complex practical problems, making GA an effective tool for solving complex optimization problems. Firstly, the electric vehicles' power system structure and energy management strategy are investigated and analyzed. Secondly, the improved non-dominated sorting genetic algorithm II (NSGA-II) is selected to optimize the parameters of electric vehicles because of its simple operation and high optimization accuracy. Thirdly, NSGA-II is used to construct the electric vehicles' power and energy configuration, with power and economic performance as the main optimization objectives. Meanwhile, a fuzzy logic controller is designed to adjust the parameters of GA online, so that the optimization process is closer to the actual operating conditions. Finally, the relevant variables are selected to achieve the optimization goal, the optimization objective function and constraint conditions are established, and the model is simulated and evaluated. The results show that the optimized electric vehicle's acceleration time is remarkably reduced, the dynamic performance is improved by more than 7 %, and the power loss is reduced by 5 %. In addition, compared with the current multi-objective optimization model, this model enables electric vehicles to travel longer distances under the same power. This study provides a new idea and method for the performance optimization of electric vehicles. Moreover, it offers a valuable reference for the innovation and development of electric vehicle technology in the intelligent manufacturing field. This study indicates that electric vehicles could be more efficient, energy-saving, and environmentally friendly to serve people's travel needs in the future.

Optimization of Ultra-High Performance Concrete Based on Response Surface Methodology and NSGA-II.

This study systematically investigated three influential factors-water-to-binder ratio, cement/sand ratio, and steel fiber content-that significantly impact the performance of ultra-high-performance concrete (UHPC). Utilizing the Response Surface Methodology (RSM) with a Central Composite Design (CCD), 20 carefully designed mix proportions underwent comprehensive experimental testing. Through rigorous statistical analysis, models were established to elucidate the complex relationships between the specified factors and the overall properties of UHPC. Variance analysis reveals significant effects of the three factors on UHPC performance, with workability and compressive strength increasing with higher cement/sand ratios while flexural strength decreases. Moreover, increased water-to-binder ratios exhibit substantial negative impacts on both 28-day compressive and flexural strengths. Despite adversely affecting workability, higher steel fiber dosages contribute positively to mechanical performance. Furthermore, Monte Carlo sampling and the multi-objective non-dominated sorting genetic algorithm-II (NSGA-II) were employed to validate the reliability of the statistical model and to conduct multi-objective optimization. The final UHPC mix design obtained consists of a cement/sand ratio of 1.12, a water/binder ratio of 0.16, and a steel fiber content of 2.94%. Experimental results yielded a slump flow of 802 mm, compressive strength of 122.7 MPa, and flexural strength of 24.3 MPa.

Evaluation of intensive remediation using simulation-optimization modeling based on long-term monitoring at a DNAPL contaminated site

Simulation-optimization modeling is extensively used to identify optimal remediation designs. However, verifying these optimal solutions often remains unclear. In this study, we determine optimal groundwater remediation strategies using simulation-optimization modeling and assess the effectiveness of previous remediation efforts by validating optimized results through 14 years of long-term monitoring of trichloroethylene (TCE) contamination. The study site is the Road Administrative Office (RAO) in Wonju, Korea, where significant TCE contamination has occurred, and long-term in-situ remediation and monitoring have been conducted. We employ MODFLOW for simulating groundwater flow and MT3D for modeling dissolved TCE concentration distribution. The Non-dominated Sorting Genetic Algorithm-II (NSGA-II) is applied to derive optimal groundwater remediation designs. Initial simulation results effectively predicted long-term TCE contamination trends and the impact of short-term in-situ remediation. Our evaluation involved comparing these optimal designs with field test outcomes, leading to the integration of continuous intensive pump-and-treat with in-situ remediation strategies. By comparing various modeling scenarios against long-term TCE contamination trends, we confirmed the effectiveness of previous remediation efforts and demonstrated that the optimal remediation design substantially minimized TCE concentrations at the main source zone. This study highlights successful strategies in historical contamination and remediation trend assessments, proposing an optimal design for pump-and-treat with reduced pumping stress to manage remaining TCE contamination at the site effectively.

The synergetic competition of cross-regional integrated energy systems with low-carbon technology heterogeneity under carbon emission trading mechanism to achieve carbon reduction target

The integration of renewable and fossil energy in a Region Integrated Energy System (RIES) is recognized as an effective measure for carbon emission reduction. The emergence of carbon emission trading (CET) markets worldwide has raised the need to combine RIES optimal planning with CET to achieve carbon reduction targets. This study focuses on the synergetic competition of a cross-RIES system with low-carbon technology heterogeneity under the CET mechanism. A cross-RIES model is constructed, incorporating fossil Carbon Capture System (fossil-CCS), renewable Combined Cooling Heating and Power (renewable-CCHP), and cleaner production paradigms. An improved version of the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) is used to propose a bi-level multi-objective optimal planning framework for the cross-RIES model, employing a grandfather permit allocation method. Comparative analysis is conducted between cross-RIES optimal planning with and without considering the CET mechanism, with a focus on environmental indexes such as carbon emission, carbon intensity, and marginal abatement cost. Sensitivity studies are also performed to explore the impact of key CET design parameters on the system's environmental performance. The simulation results demonstrate that the CET mechanism effectively reduces carbon emission, primary energy consumption, and carbon intensity by promoting synergetic competition among the low-carbon technology heterogeneity of RIESs. Furthermore, the environmental performance varies among the three types of RIES. In terms of CET parameter design, lowering the cap constraint, reducing the allowance allocation ratio of fossil-CCS RIES, and increasing the trading price contribute significantly to reducing carbon emission and carbon intensity, albeit at the expense of higher marginal abatement cost.

Enhancement of joint quality for laser welded dissimilar material cell-to-busbar joints using meta model-based multi-objective optimization

In the battery pack assembly, it is essential to ensure that the cell-to-busbar joints can be produced with high quality and with minimal impact on the individual battery cells. This study examines the influence of process parameters on the joint quality for nickel-plated copper and steel plates, laser welded in an overlap configuration. Artificial neural network-based meta models, trained on numerical results from computational fluid dynamics simulations of the laser welding process, are used to predict and evaluate the joint quality. A set of optimized process parameters is identified, in order to simultaneously maximize the interface width for the joints, and minimize the formation of undercuts and in-process temperatures. In an meta model-based multi-objective optimization approach, the non-dominated sorting genetic algorithm II (NSGA-II) is used to efficiently search for trade-off solutions and the meta models are used for objective approximation. As a result, the objective evaluation time is decreased from around 9 h, when evaluated directly from numerical simulations, to only tenths of a second. From the Pareto-optimal front of trade-off solutions, three optimal solutions are selected for validation. The selected solutions are validated through laser welding experiments and numerical simulations, resulting in joints with large interface widths and low in-process temperatures without a full penetration.

Research on Design and Staged Deployment of LEO Navigation Constellation for MEO Navigation Satellite Failure

Low Earth orbit (LEO) satellites have unique advantages in navigation because of their high signal intensity and rapid geometric changes in a short period. In order to solve the problem of constellation performance degradation after a potential failure pertaining one or more medium Earth orbit (MEO) navigation satellites, this paper designs the LEO navigation constellation and considers the task requirements of different stages of constellation deployment. Firstly, the LEO navigation constellation is designed by a non-dominated sorting genetic algorithm II (NSGA-II). The average position dilution of precision (PDOP) is 1.676, which is an improvement compared to the average PDOP offered by the four traditional GNSS. Secondly, the staged deployment of constellation takes into account the degradation of constellation performance caused by the failure of MEO navigation satellites, and the Monte Carlo method is used to analyze the case of three simultaneous satellite failures. The results show that a single satellite failure within each orbital plane and adjacent satellites with close phase separation has a great impact on the performance of the MEO navigation constellation. On this basis, a staged deployment strategy was adopted in order to balance cost, risk, and performance. The three phases deploy 66, 156, and 288 satellites, respectively; as a make-up constellation under contingencies, a navigation enhancement constellation, and an independent navigation constellation, the deployment of the staged sub-constellations meets the mission requirements. The constellation design and staged deployment method proposed in this paper can provide reference for the future study of LEO navigation constellations.

BIM-based optimization approach to reduce life cycle costs by focusing on the integration of construction and operation phases in office-commercial buildings

This study investigates the application of Building Information Modeling (BIM) to reduce life cycle costs (LCC) by addressing critical nonconformances encountered during the construction and operational phases of buildings. The primary aim is to identify nonconformances that escalate operational costs and to evaluate how BIM can effectively reduce these costs. The research begins with a comprehensive review of existing literature and utilizes the Delphi method to identify four key nonconformances that most significantly impact operational expenses. To further analyze these issues, semi-structured interviews and the Non-dominated Sorting Genetic Algorithm II (NSGA-II) are employed to determine the cost ranges of components associated with these nonconformances in both the construction and operational phases. The study finds that BIM-based optimization can lead to substantial cost reductions, although there are limitations due to the lack of precise data on the operational phase of low-cost equipment. The key contribution of this research is the development of a framework that helps designers create more effective BIM models, enhancing facilities management (FM) and reducing LCC. The originality of this study lies in its integrated approach, combining BIM and optimization techniques to directly address operational cost issues, which provides a novel perspective on improving life cycle cost management in buildings. Future research should expand the dataset and explore a broader range of nonconformances to refine and extend the framework's applicability.

Day-Ahead Nonlinear Optimization Scheduling for Industrial Park Energy Systems with Hybrid Energy Storage

Hybrid energy storage can enhance the economic performance and reliability of energy systems in industrial parks, while lowering the industrial parks’ carbon emissions and accommodating diverse load demands from users. However, most optimization research on hybrid energy storage has adopted rule-based passive-control principles, failing to fully leverage the advantages of active energy storage. To address this gap in the literature, this study develops a detailed model for an industrial park energy system with hybrid energy storage (IPES-HES), taking into account the operational characteristics of energy devices such as lithium batteries and thermal storage tanks. An active operation strategy for hybrid energy storage is proposed that uses decision variables based on hourly power outputs from the energy storage of the subsequent day. An optimization configuration model for an IPES-HES is formulated with the goals of reducing costs and lowering carbon emissions and is solved using the non-dominated sorting genetic algorithm II (NSGA-II). A method using the improved NSGA-II is developed for day-ahead nonlinear scheduling, based on configuration optimization. The research findings indicate that the system energy bill and the peak power of the IPES-HES under the optimization-based operational strategy are reduced by 181.4 USD (5.5%) and 1600.3 kW (43.7%), respectively, compared with an operation strategy based on proportional electricity storage on a typical summer day. Overall, the day-ahead nonlinear optimal scheduling method developed in this study offers guidance to fully harness the advantages of active energy storage.

Optimization of co-injecting CH4 with CO2 to enhanced oil recovery and carbon storage: A machine-learning based case study on H59 block of Jilin Oilfield, China

During CO2-EOR implementation, CH4 in the reproduced CO2-rich mixture are included in the recycle injection gas. However, improper injection gas composition and operation parameters can reduce the flooding performance.In this study, a machine-learning assisted framework combining Convolutional Neural Networks-Gate Recurrent Unit (CNN-GRU) and Non-dominated Sorting Genetic Algorithm II (NSGA-II) was proposed to obtain the optimal parameters for simultaneous maximum oil production, carbon storage efficiency, and net present value. A case study from the CO2-EOR project in the H59 block of Jilin oilfield, China, was carried out. Base cases were established to evaluate the effects of reinjection CH4 on flooding performance.Results show that the built CNN-GRU model has high prediction accuracy and computational efficiency, and thus can be used as an alternative tool to the reservoir simulator. The proposed framework can find the optimum parameters to improve oil recovery, carbon storage and net present value. The co-injection of CH4 and CO2 improves oil production, and carbon storage performance, but reduces the net present value. This work provides engineers with multiple strategies for decision-making to simultaneously promote flooding performance with CH4 being a co-injectant in CO2-EOR projects.

Development and testing of a novel vibratory compaction instrument considering its natural frequency.

This study aims to develop a novel vibratory compaction instrument designed for the preparation of large triaxial soil specimens, offering an efficient alternative to the traditional manual specimen preparation process. However, when the motor frequency closely aligns with the natural frequency of the structure, severe resonance phenomena may occur. This paper employed finite element modal analysis, experimental modal analysis, and operational modal analysis techniques to conduct a multi-objective optimization of the structure, aiming to improve the dynamic characteristics. First, the natural frequency and mode shapes of the vibratory compaction instrument were determined via finite element technology. Subsequently, the optimization of this instrument was conducted by integrating the Kriging surrogate model with the Non-dominated Sorting Genetic Algorithm II (NSGA-II). Finally, the steel column underwent a modal test using impulse stimulation to ascertain its natural frequency, and the stochastic subspace identification (SSI) approach was employed to conduct an operational modal analysis of the vibratory compaction instrument to determine its modal characteristics. The experimental findings corroborated the conclusions obtained from the finite element analysis, validating the precision of the finite element modal analysis and the rationality of the optimized structure. The research results are of significant importance for improving the preparation efficiency of large triaxial specimen, contributing to advancements in soil mechanics testing.

A trajectory optimization method for CAVs at intersections considering energy-efficiency balance

With the development of vehicle-to-everything (V2X) technology, connected automated vehicles (CAVs) will be an important part of urban road transportation. However, frequent acceleration, deceleration, and rapid stopping within intersections lead to increased energy consumption, reduced driving efficiency and balanced benefits. The study proposes an optimal control method for vehicle trajectories that considers energy-efficiency balance (OCM-EEB) based on the characteristics of V2X technology. The method first develops a vehicle trajectory control model based on the Intelligent Driver Model (IDM) by accounting for the features of V2X technology, and further, introduces the idea of balanced optimization, considering energy-efficiency balance, a vehicle trajectory optimal control model (OCM) is proposed. The proposed model was solved by design algorithms based on the non-dominated sorting genetic algorithm-II (NSGA-II). The experimental results showed that the vehicle trajectory optimization control method could increase the balanced driving efficiency of the vehicle within intersection, which is conducive to improving the robustness of the vehicle’s trajectory state, and then prove the effectiveness of the method proposed in this study. The method also provides an important technical basis for the vehicle trajectory optimization experiments under a real V2X environment in the future.

Intersecting social welfare with resilience to streamline urban flood management

Urban policymakers have long searched for stormwater management plans that incentivize stakeholders to adopt Green Infrastructure (GI) while effectively reducing the vulnerability of drainage systems. In this regard, our research introduces a novel framework to develop GI strategies that provide both hydrological resiliency and social acceptance. To achieve this, first, using a coupled Stormwater Management Model (SWMM) and Non-dominated Sorting Genetic Algorithm-II (NSGA-II), optimal alternatives for GI planning were generated. In the optimization process, we used a novel Simple Urban Flood Resilience Index (SUFRI) to consider the internal performance of the system in identifying resilient plans. Derived management strategies warrant runoff volume reduction and resilience improvement up to 31.3% and 55.1%, respectively. In the next step, Utilitarian-based Social Welfare (USW) was employed to clarify the socio-economic behavior of management strategies. Results indicate that while financial incentives significantly motivate developers to implement GI, they cannot guarantee high social welfare, and achieving a sustainable solution requires evaluating both SUFRI and USW layers under different subsidy levels. Visualizing the SUFRI layer revealed a critical failure in the resiliency trend of solutions that cannot be detected by evaluating simpler metrics, such as runoff volume reduction. This highlights the importance of the SUFRI method in conducting deeper evaluations and preventing financial waste. Finally, we navigated the intersection of USW and SUFRI measures to reach an ideal management plan with optimal supporting level. Our findings showed that the selected solution with the highest social acceptability can improve the resiliency of the system by 29 %. This study is a novel combination of the hydrological and social aspects of stormwater management, enabling decision-makers to take significant steps towards sustainable urban development.

Multi-objective optimization of building energy consumption and thermal comfort based on SVR-NSGA-II

It is necessary to reduce the huge energy consumption of buildings while ensuring indoor thermal comfort. The building envelope design notably affects the energy consumption and indoor thermal comfort. This study proposes a framework of support vector regression non-dominated sorting genetic algorithm-II (SVR-NSGA-II) to study the multi-objective optimization of buildings and discover the best building envelope design. First, the data are obtained by OpenStudio 1.1.0 simulation. Then SVR is used to describe the relationships between the parameters and the two objectives. Using SVR models as the objective functions, NSGA-II is applied to optimize the objectives. Finally, taking a proposed office building as an example, the developed framework is applied, verifying its feasibility and effectiveness. The optimized parameters of the building envelope are as follows: the exterior wall U-value, roof U-value, exterior wall U-value, solar heat gain coefficient value (SHGC), and south, north, east, west window-to-wall ratios are respectively equal to 0.664 , 0.393 , 1 , 0.381, 0.697, 0.7, 0.273, and 0.254. The design of the proposed building can be modified to provide a more energy efficient and comfortable building. The proposed model saves time and simplifies operation compared with traditional building information modeling methods.

Integrating machine learning with dynamic multi-objective optimization for real-time decision-making

Real-time decision-making in dynamic multi-objective optimization problems (DMOPs) is challenging due to constantly changing objectives and constraints. This paper integrates machine learning with Non-dominated Sorting Genetic Algorithm II (NSGA-II) to solve DMOPs and make real-time decisions. Learning-based methods have gained popularity for predicting solutions in new environments and capturing changing patterns in optimal solutions. However, existing approaches often struggle with training difficulty and reduced prediction accuracy due to irrelevant or redundant variables. Therefore, we introduce a new interdependent prediction (IDP) technique to identify correlations between variables and prediction targets and select significant variables for a predictive model. In this way, a better initial population is predicted. The IDP strategy is integrated within the dynamic NSGA-II, introducing a new algorithm called IDP-DNSGA-II. This integration facilitates rapid convergence, finding optimal or near-optimal solutions. The proposed method is evaluated against standard benchmarks, demonstrating superior performance in convergence speed and solution diversity with the changes in the problem environment. The IDP-DNSGA-II is validated through real-world optimization challenges in sustainable automobile production distribution in order-to-delivery systems to enhance environmental sustainability and operational efficiency. This study identifies the minimum frequency of change required in real-world problems to adequately track the optimal decision in real-time.