Earticle

Task offloading in multi-access edge computing (MEC) systems is critical for managing computational tasks in dynamic urban environments. Existing strategies face challenges such as high communication overheads and regional performance deviations including centralized and distributed methods. Clustering approaches have been explored to address these issues, yet they often rely on physical proximity to form clusters, overlooking the variability in task rate distributions across edges. To overcome these limitations, this paper proposes a graph-driven inter-cluster resource distribution (GIRD) clustering scheme that clusters edge nodes based on task request distribution and computing resource status, ensuring similar resource utilization across clusters. Building on this, a proximal policy optimization (PPO)-enabled intra-cluster task offloading algorithm (PITO) is introduced to determine one execution server for task offloading?either an edge server within a cluster or a cloud server?using various network state information. This dynamic decision-making process optimizes a multi-objective function that includes task processing delay, consumed energy, success rate, and cloud cost. Simulation results demonstrate the proposed GIRD-PITO framework achieves superior task success rates, reduced delays, and improved regional performance fairness, making it a promising solution for large-scale MEC systems.

In cloud computing systems the continuous increasing need for a variety of applications has made workload distribution and resource allocation more difficult. These difficulties lead to a high degree of imbalance and poor system performance. To address these issues, this study presented a high scalable Improved Multi-Objective Hybrid Load Balancing (IMH_LB) Algorithm. It combines the Grey Wolf Optimization (GWO) algorithm with the velocity driven approach of Particle Swarm Optimization (PSO) technique. The suggested method has been developed to optimize cloud computing environments by improving resource utilization with high scalability. The quality-of-service (QoS) parameters are comparatively analyzed in CloudSim by benchmarking the proposed algorithm against standard (GWO and PSO) algorithms and state of the art (Hybrid GWO-PSO and Improved Hybrid Genetic Algorithm-GWO) algorithms. The experimental findings show the superiority of the proposed approach with average improvements of 72.16% in average response time, 67.27% in makespan, 5.18 times in throughput, in 7.51 times in average resource utilization and 78.51% in degree of imbalance over the standard (GWO and PSO) and state of the art (HGWO-PSO and IHGA-GWO) methods. The proposed algorithm achieves high scalability while stabilizing at an average utilization of 89 % at significantly high workloads.

도시 내에서 발생한 화재와 같은 재난은 도시 내 광범위한 지역에서 다수의 인명, 경제적 피해를 야기한다. 이러한 피해를 최소화하기 위하여 강화학습을 활용한 효과적인 재난 대응 팀의 배치 방안이 주로 연구되었으나, 기존 방안은 재난 대응팀의 관할 지역의 크기가 증가될 때 효과적인 재난 대응을 기대하기 어렵다. 따라서, 본 연구팀은 대규모의 재난 상황에서 기존 강화학습 기반 대응팀 배치 방식의 문제를 해결하기 위해 상충되는 목표들 간의 가중치를 조절하는 새로운 강화학습 기법과 각 대응 목표들의 정의를 화재 재난 상황을 기반으로 설계하고 제안한다.

안정적인 수돗물 공급을 목적으로 하는 상수도 시설은 설계비용과 에너지 비용 등 총 비용을 저감할 수 있도록 계획되어야 한다. 상수관망 설계 시 관경에 따라 설계비용이 달라지고 마찰 손실 에너지가 달라지기 때문에 적절한 관경을 결정하는 것이 효과적인 관망 관리에 도움이 된다. 본 연구에서는 설계비 용의 최소화와 마찰 손실 에너지의 최소화를 목적으로 하는 다목적 최적화 연구를 수행하기 위해 비지배 정렬 유전자 알고리즘(NSGA-Ⅱ)과 상수관망 프로그램인 EPANET을 연계하여 적용하였다. 대상지역은 인천시 청라지역 3개의 동으로 하였고, 마찰 손실 에너지를 산출하기 위한 관 내의 유량과 마찰 손실 수두는 EPANET 모델 결과를 활용하였다. 연구를 통하여 대상지역 상수관망 내 수요량과 압력을 충족시 키면서 설계비용과 마찰 손실 에너지의 파레토 프런트를 구성하였다. 또한, 비지배 정렬 유전자를 이용하 여 두 목적함수를 만족시키는 관망의 최적 설계 연구가 가능하다는 것을 제시하였다.

Water supply system that provide safe water to customers should be designed to reduce construction cost and management cost. Pipe diameter is a key aspect of water distribution system since it affects design cost and friction loss. This study applied multi-objective optimization algorithm coupled with EPANET hydraulic solver for minimizing design cost and friction loss energy to water distribution system. Non-dominated sorting genetic algorithm (NSGA-Ⅱ) was used as an optimization technique. The hydraulic results such as head loss and flow in the pipeline were obtained from EPANET program. The test bed was the Cheongna water distribution system in Incheon, Korea. The findings in this study provided Pareto front of minimizing design cost and friction loss energy with simultaneously satisfying the water demand and hydraulic pressure. This study suggests that NSGA-Ⅱ can be used to design optimal water distribution system that satisfy the multi-objective options.

불균질 저류층에서 불규칙한 생산폐쇄가 존재할 때 개별 생산정의 생산추이를 정확하게 예측할 수 있는 다목적함수 기반의 히스토리 매칭법을 개발하였다. 개별정의 생산추이는 비지배정렬과 밀집도거리를 결합한 유전 알고리즘을 통해 최적화하였다. 최적치는 낮은 오차순으로 순위를 평가하고, 밀집도거리를 통해 다음 세대 유전자 추출의 다양성을 제공함으로써 광역 최적해에 도달하게 하였다. 각기 다른 원유 생산량을 가진 3개의 유정에서 개별정 물생산비를 관찰하였다. 개발 모델은 오차가중합법을 사용한 기존의 히스토리 매칭에 비해 불규칙한 생산폐쇄에도 25-85% 가량 향상된 예측성능을 보였다. 또한 다수 최적해를 공간적으로 분포시킨 파레토면을 통해 추가적인 민감도 분석 없이 최적해의 신뢰도를 평가할 수 있었다. 또한, 가중치를 임의로 설정하지 않아도 안정적인 예측성능을 보임으로써 가중치 설정에 따른 예측민감도를 극복하였다. 이 연구에서 개발한 유전 알고리즘 기반의 다목적 히스토리 매칭법은 다양한 생산거동을 보이는 다중정의 생산량 예측에 유용하며 추가 생산정 설계 및 생산효율성 평가에 활용 가능하다.

The paper presents a multi-objective history matching to reduce the performance uncertainties and forecast it reliably in a heterogeneous reservoir with multiple production wells. Individualized well performance is optimized separately using a genetic algorithm coupled with non-dominated sorting and diversity preservation. The fitness is sorted along to the proximity and then the diversity is added by examining the crowding distance as the approach to arrive at the global optimum. Individual watercut trajectory is observed in a heterogeneous oil reservoir with three production wells. The model depicts the effect of unexpected shut-in/out more accurately than the conventional history matching based on linearly weighted objective function. The prediction accuracy is improved up to around 25 to 85% compare to the former. Each performance shows stable prediction regardless of weights. Without extra sensitivity analysis and intuitive selection of weights, it can find out another feasible solution on Pareto front.

이 논문은 다목적 최적화 알고리듬을 이용하여 오일샌드 저류층에서 생산성이 높으면서도 최적의 에너지효 율을 갖는 SAGD(Steam Assisted Gravity Drainage) 공법의 운영조건을 선정하는 기법을 제안한다. 기존의 SAGD 공 법 최적화 연구들은 경제성인자가 고정된 단목적 최적화에 집중하였다. 이 연구에서는 회수율 최대화와 누적증기오 일비 최소화 관점에서 비지배관계에 놓이는 최적 생산운영 시나리오를 선정하였다. 이를 통하여 특정 목적함수에 편 향된 하나의 최종해만을 도출하는 기존 방법의 한계를 개선하였다. 제안한 방법으로 선정된 최적운영 시나리오들은 기존 방법과 달리 유가상황 및 비용 변화에 따라 추가적인 최적화 작업 없이 광구운영 의사결정에 유용하게 활용할 수 있다.

This paper proposes multi-objective optimization of the Steam Assisted Gravity Drainage (SAGD) process for improving the energy efficiency and recovery factor of oil sand reservoirs. Previous studies conducted on optimizing the SAGD process have focused on single-objective optimization with fixed economic factors. In this study, multiple trade-off operating scenarios were selected by applying a multi-objective optimization algorithm that aims at maximizing the recovery factor and minimizing the cumulative steam-oil ratio for efficiently addressing volatile market conditions. Thus, the proposed method can overcome the limitations of conventional optimization methods that not only yield a single solution based on a particular objective function but also are hard to adapt to fluctuating oil prices. Furthermore, the proposed method can provide optimum trade-off operating scenarios, and hence can aid in planning operating (i.e., marketing) strategies according to the variation in oil prices and operating costs, without the need for an additional optimization process.

이 연구는 CO2 수송-주입-저장 통합 시스템의 기본 설계 단계에서 다목적 최적화에 기반한 설비및 운영 인자 선정 방법을 제안한다. CO2 지중저장 통합 시스템의 설계 인자에 대한 다목적 최적화를 수행하여 최적의 비지배 해집합인 파레토 최적면을 탐색함으로써 상호 비지배 관계인 등가의설계안을 다수 도출한다. 방법론의 성능은 대염수층 대상 CO2 지중저장 통합 시스템의 기본 설계에 적용하여 검증하였다. 결정변수는 CO2 송출 압력, 수송관 내경, 주입정 튜빙 내경, 주입정 상단에서의 CO2 온도 4종류를 설정하였고, 목적함수는 CO2 저장량, 주입정 공저압, 편익비용비 3종류를 설정하였다. 연구 결과, 전체 탐색공간 대비 약 5% 수준의 계산 비용으로 파레토 최적면을 도출하여 설계 인자 범위를 확인하였다. 또한, 최적화 목적함수에 대한 사전 선호도와 사후 선호도 적용영향을 비교 분석하였다. 제안하는 방법론은 향후 CO2 지중저장 통합 시스템 설계 과정에서 검토할 수 있는 우수해를 다수 제공할 수 있음을 확인하였다.

This study proposes a method based on multi-objective optimization to select facility and operational parameters for the basic design of an integrated CO2 transportation-injection-storage system. Ｍulti- objective optimization was conducted to explore a Pareto-optimal front composed of non-dominated design solutions. The performance of the proposed approach was validated by applying it to CO2 storage at a saline aquifer. We set four decision variables (CO2 source pressure, pipeline inner diameter, injection well tubing inner diameter, and CO2 temperature at wellhead) and three objective functions (CO2 storage capacity, injection pressure, and benefit-cost ratio). The Pareto-optimal front was derived at approximately 5% of the computational cost compared to the exhaustive search. A comparative analysis was conducted to analyze the effects of prior and posterior preferences toward objectives on optimization performance. This method demonstrates the potential to provide effective solutions for designing an integrated CCS (carbon capture and storage) system.

국내는 정보통신(IT) 기술의 발달을 기반으로 BIS, BMS, CNS 등을 통해 교통정보가 실시간적으로 제공되고 있어 유비쿼터스 기반의 첨단 통신기술이 정착되게 되면 개인의 취향에 적합한 맞춤형 (대중)교통정보가 가능할 것으로 전망된다. 대중교통 통행정보는 경로탐색알고리즘으로 구현되는데, 대중교통망은 일반 가로망과는 달리 환승과 운행시간 특성을 포함하며 통행자가 경로를 선택하는 과정에서 통행시간, 환승횟수, 환승시간과 같은 복수의 제약 또는 목적의 고려가 요구된다. 본 연구에서는 사용자에게 영향을 미칠 수 있는 요인들을 복수의 목적으로 고려하여 경로를 탐색하며 탐색된 경로에 대하여 자의적으로 선택해 나가는 기법을 제안한다 본 연구는 복합대중교통망을 대상으로 통행시간, 환승횟수, 환승시간등과 같은 통행목적에 대하여 계층별로 별도의 다목적 경로를 탐색하는 방안을 강구하여 계층별로 적용 가능한 사용자 맞춤형 정보제공체계의 기본적인 접근이론을 개발한다. 제안된 다수단, 다계층, 다목적 경로탐색알고리즘은 링크표지에 근거하여 구축되어 복합교통망에 효율적으로 구동할 수 있으며, 수도권 도시철도의 실제 대규모 교통망에 적용한 사례연구를 통하여 다양한 계층을 고려한 경로계획모형(Path Planning)으로 활용될 수 있는 가능성을 예시하였다.

Mass transit information should contribute many benefits to users. Especially transportation information technology is developing highly with information technology in Korea recently. Hereafter it is expected to provide customized transportation information to users individually with the advent of ubiquitous age in earnest. This public transportation information service can be realized by path finding algorithms in public transportation networks including travel and transfer attributes. In this research, multi objectives such as travel time, transfer time, and number of transfer and so on are constructed with the primary facts influencing users. Moreover, the method reducing user's path finding alternatives arbitrarily is proposed by selecting the best alternative which provides maximum utility to users among non dominated paths. Therefore, the ultimate goal of this study proposes a multi objective shortest paths finding algorithm which can take into account multiple user classes in a transit network with multiple travel modes. The proposed algorithm is demonstrated based on the two case studies - a small toy network and the large-scaled Seoul Metropolitan subway network.

According to Genetic algorithms principle, the new hybrid evolutionary algorithm (HEA) is proposed in this paper by combining the Immune algorithm, Genetic algorithm and Pareto optimal solutions. The HEA has high convergence precision and improved the diversity of population. Multiple near optimization paths can be developed by the algorithm with multi‐objective restriction, and satisfy to minimize the routing of transportation and the numbers of the vehicles. The HEA has been used to solve the vehicle routing problem, the results of simulation experiment show that the HEA can gain higher global convergence rate and higher speed.

In various machines used in industrial sites and transportation equipment, fastening structures of bolts and nuts are widely employed. However, conventional Steel sockets, classified as non-explosion-proof materials, have a high likelihood of generating sparks due to friction with components, which can lead to explosions or large-scale fires. To address this issue, this study developed a lightweight explosion-protection socket using AL-7075-T6 aluminum alloy, which is known for its excellent explosion-proof properties. However, due to the inherent characteristics of aluminum, it has lower rigidity compared to Steel, requiring the use of more expensive alloy materials. Therefore, our research team utilized Finite Element Analysis (FEA) and Multi-Objective Genetic Algorithm (MOGA) to optimize the mass and safety factor of the socket, proposing a design that simultaneously achieves both weight reduction and structural stability. The socket developed in this study is approximately 30% lighter than traditional Steel-based sockets while maintaining a safety factor of 1.2 or higher, significantly enhancing operational safety in explosive environments. This research sets a new standard in the design and manufacturing process of explosion-proof sockets and is expected to contribute to the optimization of various explosion-proof equipment in the future.

본 연구에서는 다중최적화기법을 이용하여 2가지 수문학적 과정을 통하여 유출량을 산정하는 수문모형의 모형 최적화를 시도하였으며, 수문모형으로는 융설량과 유출량을 동시에 산정할 수 있는 분포형 수문모형인 HL-RDHM을 이용하였다. 대상유역으로는 융설량 자료를 수집할 수 있는 미국 콜로라도의 Durango River 유역을 선정하였다. 다중최적화기법으로는 MOSCEM을 활용하였으며, 융설과 관련된 매개변수 5개와 유출에 관련된 매개변수 13개를 선정하여 매개변수 보정과 수문모형 최적화를 시도하였다. 모형 최적화를 위해 2004 – 2005년의 자료가 활용되었고, 2001 – 2004년 자료를 이용하여 검증하였다. 융설량과 유출량을 동시에 최적화함으로써 RMSE 기준으로, 3개의 SNOTEL 지점에서 초기해에 의한 모의치 보다 7% - 40%까지 RMSE 오차를 줄일 수 있었고, 유출구의 USGS 관측점에서 초기해에 비해 약 40% 값이 개선됨을 확인하였다.

In this study, the multi-objective optimization method is attemped to optimize the hydrological model to estimate the runoff through two hydrological processes. HL-RDHM, a distributed hydrological model that can simultaneously estimate the amount of snowfall and runoff, was used as the distributed hydrological model. The Durango River basin in Colorado, USA, was selected as the watershed. MOSCEM was used as a multi-objective optimization method and parameter calibration and hydrologic model optimization were tried by selecting 5 parameters related to snow melting and 13 parameters related to runoff. Data from 2004 to 2005 were used to optimize the model and verified using data from 2001 to 2004. By optimizing both the amount of snow and the amount of runoff, the RMSE error can be reduced from 7% to 40% of the simulation value based on the initial solution at three SNOTEL points based on the RMSE. The USGS observation point of the outflow is improved about 40%.

The successful implementation of remanufacturing supply chain not only needs foundation engineering technology, but also needs the efficient supply chain model to support logistics network. But till now, the optimization problems of supply chain logistics network focus on the determination for the number and location of facilities and logistics distribution between the various facilities. Except for the decisive factors, the optimization problem of remanufacturing supply chain logistics network considers the environmental pollution factors the waste product returns and transportation. The market demand and waste product returns are actually uncertain in remanufacturing supply chain. There is few papers focus on dual uncertain factors although there are lots of studies on the problems. Therefore, based on dual stochastic programming, the optimization model of multi-phase multi-objective remanufacturing supply chain is established with maximum profit in the remanufacturing supply chain, maximum rapid response customer satisfaction and minimum environmental pollution. Dual-layer genetic algorithm mechanism was brought up. The first layer algorithm is responsible for supply chain logistics network structure. The second layer algorithm determines specific distributions for remanufacturing supply chain, based on optimized logistics network structure mechanism in the first layer algorithm. Finally, numerical examples demonstrate the validity of the model and algorithm for the optimization problem.

This paper proposes a Modified Differential Evolution (MDE) for solving multi-objective load dispatch (MOLD) problem where transmission power losses are considered. MDE is an improved version of conventional Differential Evolution (CDE) in which the mutation operation of the CDE is improved by using five differential solutions instead of three ones similar to CDE. In the MOLD problem, three cases of dispatch including economic dispatch, emission dispatch and multi-objective dispatch are carried out by considering fuel cost function, emission function and both fuel cost and emission functions. In the third case of dispatch, there is a price penalty factor employed to determine the best compromise solution instead of using Fuzzy-based mechanism similar to other studies. The performance of MDE is verified by testing on two systems with three units and one system with six units. In the two systems, the fuel cost and emission from MDE are compared to those from CDE and other existing meta-heuristic algorithms, and the analysis on the result comparison indicates that the MDE is a promising algorithm for solving the MOLD problem.

To effectively optimize multi-objective logistics distribution path, the distance and distance related customer satisfaction factor are used as the objective function, a novel kruskal crossover genetic algorithm (KCGA) for multi-objective logistics distribution path optimization is proposed. To test the optimization results, the terminal distribution model and the virtual logistics system operating model are built. Experiment results show that, compared with basic genetic algorithm (GA), the run time of KCGA takes a slightly higher. But the average distribution distance and the best distribution distance are reduced by 6%-8%. Achieve the goal of multi-objective logistics distribution path optimization.

By the point of view of remanufacturing enterprise, the studies of supply chain logistics network optimization problem focused on network optimization and logistics distribution between manufacturing enterprise and customer area. The present business environment is full of uncertainties. To compete effectively in such an environment, remanufacturing enterprises need to develop the capability of responding flexibly to changing market conditions and optimize effectively logistics networks. The mechanism of option contracts has effectively coordinated the relationship of remanufacturing enterprise and retailer and hedged against the risks of over- and under-production. There is few papers considers the mechanism of option contracts although there are lots of studies on supply chain logistics networks optimization problem. Therefore, the mechanism of option contracts is led into the study of remanufacturing supply chain logistics networks optimization problem. Considering environmental pollution influence factors in the remanufacturing process, a multi-objective optimization model of supply chain is developed. Dual-layer genetic algorithm mechanism was brought up. The first layer is responsible for supply chain logistics network structure with genetic algorithm. The second layer answers for specific supply chain logistics distribution with adaptive immune genetic algorithm. Finally, numerical examples demonstrate the validity of the model and algorithm for the optimization problem.

Considering the complexity of flexible workshop scheduling, combined with plant production process characteristics and constraints, we constructed a multi-agent system model to solve multi-objective flexible workshop scheduling problems. This paper proposed an algorithm which was a combination of the ant colony algorithm and Q-learning algorithm. This paper also analyzed and implemented how to solve the workshop scheduling optimization problem. Finally, this paper proved the validity of methods to solve the multi-objective flexible workshop scheduling optimization problems with examples on JADE platform.