simulated annealing algorithm

Developing and optimizing effective inventory systems considering realistic constraints and practical assumptions can help managers remarkably decrease inventory and consequently supply chain costs. In this research, we propose a new variant of the multi-item inventory model taking into account warehouse capacity, on-hand budget constraints, imperfect products in supply deliveries and partial backordering where the products can be converted into perfect products by a local repair shop. To deal with the proposed model, three solution approaches, including interior-point technique, as an exact method, and two metaheuristics based on Simulated Annealing (SA) and Water Cycle Algorithm (WCA), are proposed. Extensive computational experiments are conducted on different sets of instances. Using different measures such as RPD, PRE, and computational time, the performance of the solution approaches is evaluated within different test instances. The results show that the WCA outperforms the two other approaches and leads to the best solutions in the proposed problem.

Digital marketing has become vital to businesses' marketing strategies in today's technology and social media era. However, the effectiveness of digital marketing campaigns largely depends on accurately identifying the target audience. This study aims to implement the simulated annealing initiative algorithm for digital marketing, as well as audience classification and optimum target audience selection. Traditional methods of target audience identification, such as demographic, geographic, and psychographic segmentation, are only sometimes effective in identifying the most responsive audience. Therefore, advanced techniques such as clustering, genetic, and simulated annealing algorithms have been proposed to identify the optimum target audience. The heuristic simulated annealing algorithm is one of the most promising techniques for optimum target audience identification. It is widely used in combinatorial optimization problems and applied in various fields such as engineering, economics, management, and computer science. In this research, a digital marketing campaign is implemented for a new line to sell training courses in empowerment and competency in human resource management within the mining industry. After conducting market research, we have identified five critical segments: age, gender, income group, place of residence, and level of university education. The number of customers at each customer journey stage was 740 people in brand development, email, and advertising campaigns, of which 620 people are in the "Awareness" stage, 431 people in the "Interest" stage, 261 people in the "Consideration" stage, 203 people in the "Intend" stage, 179 people in the "Purchase" and finally, 179 People were evaluated in the "loyalty" stage for the case of educational service company. The results show we should target 20% of our marketing efforts towards the 18-24 age group, 30% towards females, 20% towards high-income individuals, 10% towards rural areas, and 20% towards University education level in BSc. The best cost per conversion we obtain is 78.105×106 Rials. The results show that the simulated annealing algorithm can be valuable for identifying the optimum target audience in digital marketing campaigns. By considering the entire customer journey and allowing for more complex audience targeting, the algorithm can help companies optimize their marketing strategies and maximize their profits.

The cycle basis of a graph arises in a wide range of engineering problems and has a variety of applications. Minimal and optimal cycle bases reduce the time and memory required for most of such applications. One of the important applications of cycle basis in civil engineering is its use in the force method to frame analysis to generate sparse flexibility matrices, which is needed for optimal analysis. In this paper, the simulated annealing algorithm has been employed to form suboptimal cycle basis. The simulated annealing algorithm works by using local search generating neighbor solution, and also escapes local optima by accepting worse solutions. The results show that this algorithm can be used to generate suboptimal and subminimal cycle bases. Compared to the existing heuristic algorithms, it provides better results. One of the advantages of this algorithm is its simplicity and its ease for implementation.

Nowadays, the citrus supply chain has been motivated by both industrial practitioners and researchers due to several real-world applications. This study considers a four-echelon citrus supply chain, consisting of gardeners, distribution centers, citrus storage, and fruit market. A Mixed Integer Non-Linear Programming (MINLP) model is formulated, which seeks to minimize the total cost and maximize the profit of the Citrus supply chain network. Due to the complexity of the model when considering large-scale samples, two well-known meta-heuristic algorithms such as Ant Colony Optimization (ACO) and Simulated Annealing (SA) algorithms have been utilized. Additionally, a new multi-objective ACO algorithm based on a set of non-dominated solutions form the Pareto frontier developed to solve the mathematical model. An extensive comparison based on different measurements analyzed to find a performance solution for the developed problem in the three sizes (small, medium, and large-scale). Finally, the various outcomes of numerical experiments indicate that the MOACO algorithm is more reliable than other algorithms.

In this paper, an integer linear programming formulation is developed for a novel fuzzy multi-period multi-depot vehicle routing problem. The novelty belongs to both the model and the solution methodology. In the proposed model, vehicles are not forced to return to their starting depots. The fuzzy problem is transformed into a mixed-integer programming problem by applying credibility measure whose optimal solution is an (α,β)-credibility optimal solution to the fuzzy problem. To solve the problem, a hybrid genetic-simulated annealing-auction algorithm (HGSA), empowered by a modern simulated annealing cooling schedule function, is developed. Finally, the efficiency of the algorithm is illustrated by employing a variety of test problems and benchmark examples. The obtained results showed that the algorithm provides satisfactory results in terms of different performance criteria.

Prediction of students' academic performance is essential for systems emphasizing students' greater success. The results can largely lead to increase in the quality of the educating and learning. Through the application of data mining, useful and innovative patterns can be extracted from the educational data.

In this paper, a new metaheuristic algorithm, combination of simulated annealing and genetic algorithms, is proposed for predicting students’ academic performance in educational data mining. Although metaheuristic algorithms are one of the best options for discovering the hidden relationships between data in data science, they do not separately perform well in accurate prediction of students’ academic performance. Therefore, the proposed method integrates the advantages of both genetic and simulated annealing algorithms. The genetic algorithm is applied to explore new solutions, while simulated annealing is used to increase the exploitation power. By using this combination, the proposed algorithm has been able to predict the students’ academic performance with high accuracy.

The efficiency of the proposed algorithm is evaluated on five different educational data sets, including two data sets of students of Shahid Rajaee University of Tehran and three online educational data sets. Our experimental results show  and   accuracy improvement of the proposed algorithm in comparison to the four similar metaheuristic and five popular classification methods respectively. The author(s). This is an open access article distributed under the terms of the Creative Commons Attribution (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, as long as the original authors and source are cited. No permission is required from the authors or the publishers.

As a real time process, in tuning the coefficients of PID controllers in AVRs, accuracy vs. speed is an important issue. Considering complexity of the problem and real systems requirements, various methods, including exact methods and approximation algorithms, have been implemented for this purpose. Since the conventional methods based on meta-heuristic algorithms solving this problem, generally use population-based algorithms such as GA and PSO, this paper aims to investigate the efficiency and performanceof single-solution based metaheuristics to solve this problem. So Simulated Annealing (SA) algorithm is proposed, and implemented for optimizing PID coefficients. In addition, an extension of SA is presented improving the search strategy based on neighborhood adjustment. The results indicate that the proposed algorithms as single based metaheuristics, have a good or even better performance vs. population based metaheuristics, in spite of simplicity in implementation and less computation requirements. Thisfact implies that the landscape complexity of these problems does notnecessarily require population-based algorithms. The presented method is also applied to multipleobjective functions regarding different time response criteria in output voltage and leads to better results in less time.

In this research, we present a mathematical model for allocating people to different jobs and shifting employees between related jobs. This action will reduce the repetitive activities workload and ergonomic risks at the planned time horizon, and finally increases the organization's efficiency. In this proposed model, the devices are semi-automatic and it is possible to allocated more than one task to one person. Regarding the modeling and the case study of the constraints, it is shown that the complexity of this problem type is NP-Hard, and the result of accurate methods for solving the problem is not possible in a reasonable time. Due to this Simulated Annealing (SA) algorithm is used to study the proposed model and comparison of the results of SA algorithm with the results of precise optimization methods shows the better performance of the Simulated Annealing algorithm in terms of the time and answer quality.

A bi-objective mathematical model is developed to simultaneously consider the quay crane and yard truck scheduling problems at container terminals. Main real-world assumptions, such as quay cranes with non-crossing constraints, quay cranes’ safety margins and precedence constraints are considered in this model. This integrated approach leads to better efficiency and productivity at container terminals. Based on numerical experiments, the proposed mathematical model is effective for solving small-sized instances. Two versions of the simulated annealing algorithm are developed to heuristically solve the large-sized instances. Considering the allocation of trucks as a grouping problem, a grouping version of the simulated annealing algorithm is proposed. Effectiveness of the presented algorithms is compared to the optimal results of the mathematical model on small-sized problems. Moreover, the performances of the proposed algorithms on large-sized instances are compared with each other and the numerical results revealed that the grouping version of simulated annealing algorithm outperformed simulated annealing algorithm. Based on numerical investigations, there is a trade-off between the tasks’ completion time and the cost of utilizing more trucks. Moreover increasing the number of YTs leads to better outcomes than increasing the number of QCs. Besides two-cycle strategy and using dynamic assignment of yard truck to quay cranes leads to faster loading and unloading procedure.

Due to many damages that human activities have imposed on the environment, authorities, manufacturers, and industry owners have taken into account the development of supply chain more than ever. One of the most influential and essential human activities in the supply chain are transportation which green vehicles such as electric vehicles (EVs) are expected to generate higher economic and environmental impact. To this end, designing efficient routing scheme for the fleet of EVs is significant. A remarkable issue about EVs is their need to stations for charging their battery. Due to the existence of time limitations, more attention should be paid to time spent at the charging station, so considering the queuing system at charging stations makes more precise time calculations. Furthermore, multi-graphs are more consistent with the characteristics of the transportation network. Hence, we consider alternative paths including two criterion cost and energy consumption in the network. First, we develop a mixed integer linear programming for the electric vehicle routing problem on a multi-graph with the queue in charging stations to minimize the traveling and charging costs. Since the proposed problem is NP-hard in a strong sense, we provide a simulated annealing algorithm to search the solution space efficiently and explore a large neighborhood within short computational time. The efficiency of the model is investigated with numerical and illustrative examples. Then the sensitivity analysis is performed on the proposed model to indicate the importance of the queuing system and the impact of battery capacity on the penetration of EVs.

Manning roughness coefficient is one of the most important parameters in designing water conveyance structures. Unsuitable selection of this coefficient brings up some mistakes. This research aims to present a method to determine the Manning roughness coefficient based on a combination of optimization algorithm of simulated annealing (SA) with gradually varied flow equations. Therefore, in a lab rectangular flume of 12 m, 60 cm and 65 cm in length, width and height with fixed channel bed slope of 0.0002, nine series of water level profiles were carried out. Then, an objective function based on observed and calculated water level gradient was defined to decide on manning roughness coefficient while it was minimized with simulated annealing optimization method. The values of objective function parameters were discussed by sensitivity analysis and the most optimal objective function was obtained. To measure the accuracy of coefficient obtained, Statistics indices of R2, Root mean square error (RMSE), Mean bias error (MBE), d were used. The results showed that manning roughness coefficient has a great accuracy.

In this paper, a new multi-objective time-cost constrained resource availability cost problem is proposed. The mathematical model is aimed to minimize resource availability cost by considering net present value of resource prices in order to evaluate the economic aspects of project to maximize the quality of project's resources to satisfy the expectations of stakeholders and to minimize the variation of resource usage during project. Since the problem is NP-hard, to deal with the problem a simulated annealing approach is applied, also to validate our results GAMS software is used in small size test problems. Due to the dependency of SA algorithm to its initial parameters a taghuchi method is used to find the best possible SA parameters combinations to reach near optimum solutions in large size problems.

This paper presents a new robust mathematical model for the multi-product capacitated single allocation hub location problem with maximum covering radius. The objective function of the proposed model minimizes the cost of establishing hubs, the expected cost of preparing hubs for handling products, shipping and transportation in all scenarios, and the cost variations over different scenarios. In the proposed model, a single product of a single node cannot be allocated to more than one hub, but different products of one node can be allocated to different hubs. Also, a product can be allocated to a hub only if equipment related to that product is installed on that hub. Considering the NP-Hard complexity of this problem, a GA-based meta-heuristic algorithm is developed to solve the large scale variants of the problem. To evaluate the performance of the proposed algorithm, its results are compared with the results of exact method and simulated annealing algorithm. These results show the good performance of the proposed algorithm.