Simulation-based planning and optimization in multi-echelon supply chains

Abstract

In this paper we present a methodology and simulation environment for solving multi-echelon supply chain planning and optimization problems for industries with batch and semi-batch processes. The introduced methodology is aimed to analyze efficiency of a specific planning policy over the product life cycle within the entire supply chain for automated switching from a non-cyclic to cyclic and to optimize the cyclic planning policy for products at the maturity phase. For optimization of a multi-echelon cyclic schedule, the simulation optimization algorithm developed is based on integration of the multi-objective genetic algorithm (GA) and response surface-based local search to improve GA solutions. The comparative analysis of planning policies is based on estimation of the difference between mean values of their total costs by using the Paired-t confidence interval method and evaluation of an additional cost of the cyclic schedule. The simulation environment allows one to describe input data to build the supply chain network and store it in an external file, computing effective planning policies, automatically generating and running a network simulation model, generating production rules for switching from one planning policy to another and optimizing parameters of a multi-echelon cyclic schedule. Finally, a business case is described that illustrates the practical application of the presented methodology.

Simulation-based optimization for design parameter exploration in hybrid system: a defense system example

“…Persson et al 40 integrated a neural network and a hill-climbing algorithm, while Ghatee and Hashemi 41 combined a neural network with a particle swarm optimization technique. April et al, 42 Persson et al, 43,44 and Syberfeldt et al 45 proposed a combination of a neural network and the GA. Merkuryeva et al 46 guided the search through the GA and improved GA solutions by using a response surface model (RSM)-based linear search.…”

Advantages and Disadvantages of Professional and Free Software for Teaching Statistics

“…12), used by the German industrial partner of the ECLIPS project, is given in [22]. Practical implementation of the developed supply chain management approach, based on the cyclic planning of supply chain processes, confirmed its efficiency for industries with batch and semi-batch processes, under a relatively stable demand pattern.…”

The New Generation of Operations Research Methods in Supply Chain Optimization: A Review

“…Parameter generation is the latter purpose of integrating simulation under this sub-category of simulation-based optimization frameworks that aim to mitigate the number of unnecessary assumptions in the analytic part, which in turn improves the quality and reliability of the optimum solutions. G. Merkuryeva et al [126] proposed an optimization framework for the design of multi-echelon SCN, where the simulation part generates input data, to be stored and transferred to the optimization part via VBA (Visual Basic for Applications). This framework also allows one to easily switch from one policy to another in a user-friendly simulation environment.…”

“…In another possible approach, the given social and environmental restrictions can be fulfilled in the simulation part of S-O frameworks, while the cost criterion can be optimized through an external optimizer. Other studies have optimized the customer service level in addition to the variable costs [115,117,118,126,130]. In a simulation optimization framework developed by [143], the fitness function is calculated using a lexicographic structure where the individual optimum solutions are ranked based on total cost, customer service level and the average work in progress material.…”