Optimisation of multi-stage supply chain systems by integrated simulation-variable neighbourhood search algorithm

This research aims to develop a simulation approach based on system dynamics modelling (SDM) and adaptive network based fuzzy inference system (ANFIS) for quantifying and reducing the bullwhip effect in a multi-product, multi-stage supply chain. The proposed model is comprised of three groups of variables influencing the bullwhip effect, namely the structure of a supply chain network, supply chain contributions (ordering process in regular situation or when a supplier has a promotion or shortage gaming) and supply chain performances (the number of defects and ordering lead time). As a result, a two layer simulation model is developed with three generic models. The flexibility of this proposed approach is its ability to model various types of ordering policies which are basic inventory policies, material requirement planning (MRP) system and just in time (JIT) approach. The supply chain of a beverage company was selected to validate and demonstrate the flexibility of the proposed model. The findings of the proposed simulation model are consistent with the results obtained from the case study. The error magnitude of the bullwhip effect level varied between 0 and 9% resulting in bullwhip effect reductions of up to 92%. Accordingly, the bullwhip effect levels are significantly decreased by using the proposed simulation model.

Auxiliary services in substations are fundamental systems for the operation and coordination of power electrical systems. Because of their importance, backup systems are utilized to support critical loads during main power supply failures. Hybrid renewable backup systems, comprising renewable generation and batteries, offer a more sustainable alternative compared to the current use of fossil fuel generators in substations. However, sizing these systems can be a complex task due to the uncertainty associated with interruptions and renewable source production. This paper proposes an optimization model to size hybrid renewable energy systems for auxiliary services in substations. Uncertainties related to wind and photovoltaic generation, as well as power outages start time and durations, are addressed through Monte Carlo simulations. Furthermore, a hybrid algorithm, based on Genetic Algorithm (GA) and Variable Neighborhood Search (VNS), is introduced. The proposed approach considers the costs associated with diesel generator usage and instances of non-supply. A comparison of the metaheuristics GA, VNS, and the new hybrid algorithm is presented, emphasizing the advantages of the hybrid approach. The algorithms are validated by comparing them with results from the literature, demonstrating their ability to achieve solutions with high accuracy and low computational time. The results demonstrate that the proposed sizing approach successfully minimizes costs while ensuring reliability, highlighting its potential to optimize hybrid renewable backup systems for critical substation loads.

Optimal models for a multi-stage supply chain system controlled by kanban under just-in-time philosophy

Adaptive variable neighborhood search for solving multi-objective facility layout problems with unequal area facilities

In this paper, we study the multiobjective version of the set covering problem. To our knowledge, this problem has only been addressed in two papers before, and with two objectives and heuristic methods. We propose a new heuristic, based on the two-phase Pareto local search, with the aim of generating a good approximation of the Pareto efficient solutions. In the first phase of this method, the supported efficient solutions or a good approximation of these solutions is generated. Then, a neighborhood embedded in the Pareto local search is applied to generate non-supported efficient solutions. In order to get high quality results, two elaborate local search techniques are considered: a large neighborhood search and a variable neighborhood search. We intensively study the parameters of these two techniques. We compare our results with state-of-the-art results and we show that with our method, better results are obtained for different indicators.

The aim of the work is to create a scenario modeling technique based on adaptation of metaheuristic multi-objective optimization algorithms to construct Pareto-optimal investment portfolios taking into account multiple restrictions on the company's portfolio. Investment portfolio is a tool for company's resources management. The process of portfolio construction through a combination of individual investment projects aims to optimize available resources utilization. Currently, there is no single approach to determining the optimal portfolio, since portfolio requirements depend on a particular company. Due to its simplicity in terms of implementation, approaches based on the ranking of projects according to individual technical and economic indicators are widely used. However, if a company needs to take into account several indicators, then this approach is not applicable. An effective solution to this problem is to create a technique that allows to find Pareto-optimal portfolios. In this work, the problem statement and analysis of existing solving approaches were carried out. Based on that, methodology of scenario modeling of a company's portfolio was created. The posed problem belongs to the class of NP-hard problems and for its efficient solving it is necessary to use metaheuristic algorithms. These algorithms include genetic algorithms, variable neighborhood search, bees algorithm, etc. Genetic algorithm NSGA-II is widely used for solving multi-objective optimization problems due to its high performance and ease of use. This algorithm was adapted to solve proposed problem. Based on it, a methodology was developed for the formation of Pareto-optimal investment portfolios, taking into account financial and infrastructural restrictions. Testing of the methodology was carried out on a dataset of synthetic projects simulating real investment projects. Various combinations of this projects simulated investment portfolios. The optimization objectives were maximization of net present value (NPV), maximization of cumulative oil production and minimization of total investments. In order to control the quality of the results obtained using genetic algorithm, the problem was also solved by using variable neighborhood search. However, the quality of the results obtained using genetic algorithm was higher. Analysis of the results allows us to consider the use of multi-objective optimization as a tool for forming a pool of investment portfolios, which provides flexible company management. In the proposed work, it is shown that the problem of scenario-based modeling of company's investment portfolio can be considered as a multi-objective optimization problem. Based on the existing metaheuristic algorithms, a Pareto-optimal investment portfolio search methodology has been developed. Proposed methodology can effectively solve the problem of managing company resources.

A heuristic-based hybrid genetic-variable neighborhood search algorithm for task scheduling in heterogeneous multiprocessor system

Artificial chromosomes with genetic algorithm 2 (ACGA2) for single machine scheduling problems with sequence-dependent setup times

The problem of selecting sugarcane varieties has been widely discussed due to its computational complexity and its great impact for the sugar and ethanol industry. This paper proposes a new integrated mathematical programming model to deal with the selection of sugarcane varieties to be planted and the determination of the optimal period for planting and harvesting in order to increase production in the sugarcane industry. The proposed model optimizes the production of sugarcane and improves the quality of biomass whilst satisfying the main constraints imposed by sugarcane companies. The problem is modelled as an integer linear program and solved using an exact method to generate optimal solutions for small and medium problems. For large problems, metaheuristic approaches based on Genetic Algorithm and Variable Neighbourhood Search are proposed. According to the results, the proposed methodology provides sugarcane company managers with decision support in selecting the most suitable varieties and in determining the best period to plant and harvest their sugarcane.

Production scheduling problems are typically combinational optimization problems named bases on the processing routes of jobs on different machines. In this paper, the flexible job shop scheduling problem aimed to minimize the maximum completion times of operations or makespan is considered. To solve such an NP-hard problem, variable neighborhood search (VNS) based on genetic algorithm is proposed to enhance the search ability and to balance the intensification and diversification. VNS algorithm has shown excellent capability of local search with systematic neighborhood search structures. External library is improved to save the optimal or near optimal solutions during the iterative process, and when the objective value of the optimal solutions are the same, the scheduling Gantt charts need to be considered. To evaluate the performance of our proposed algorithm, benchmark instances in different sizes are optimized. Consequently, the computational results and comparisons illustrate that the proposed algorithm is efficiency and effectiveness.

Redundancy allocation of series-parallel systems using a variable neighborhood search algorithm

In this paper we present two recent metaheuristics, particle swarm optimization and differential evolution algorithms, to solve the single machine total weighted tardiness problem, which is a typical discrete combinatorial optimization problem. Most of the literature on both algorithms is concerned with continuous optimization problems, while a few deal with discrete combinatorial optimization problems. A heuristic rule, the smallest position value (SPV) rule, borrowed from the random key representation in genetic algorithms, is developed to enable the continuous particle swarm optimization and differential evolution algorithms to be applied to all permutation types of discrete combinatorial optimization problems. The performance of these two recent population based algorithms is evaluated on widely used benchmarks from the OR library. The computational results show that both algorithms show promise in solving permutation problems. In addition, a simple but very efficient local search method based on the variable neighbourhood search (VNS) is embedded in both algorithms to improve the solution quality and the computational efficiency. Ultimately, all the best known or optimal solutions of instances are found by the VNS version of both algorithms.

Designing a sustainable supply chain network integrated with vehicle routing: A comparison of hybrid swarm intelligence metaheuristics

Improved Parameterization of the Transcranial Doppler Signal

A controlled migration genetic algorithm operator for hardware-in-the-loop experimentation