Buffer allocation design for unreliable production lines using genetic algorithm and finite perturbation analysis

Abstract

The buffer allocation problem in production lines is an NP-hard combinatorial optimisation problem. This paper proposes a new hybrid optimisation approach (using simulation) relying on genetic algorithm (GA) and finite perturbation analysis (FPA). Unlike the infinitesimal perturbation analysis, which deals with small (infinitesimal variation) perturbations for estimating gradients of the performance measure, FPA deals with larger (finite) or more lasting perturbations. It is an extension specifically dedicated to discrete decision variables and applicable to most discrete-event dynamic systems. The proposed method allows a global search using GA, with refinement in specific solution-space regions using FPA. The main objective is to maximise the average production rate of a production line with unreliable machines, by allocating the total buffer capacity in locations between machines. Extensive numerical experiments show that: (1) the proposed hybrid GA-FPA method clearly outperforms the state-of-the-art methods from the literature; (2) combining FPA and GA is beneficial when compared to employing GA or FPA independently.

Joint Allocation of Buffer Sizes and Service Times in Unreliable Assembly-Disassembly Systems

“…The first step consists to generate the initial population, 𝑃 = (𝑟 1 , 𝑟 2 ,...,𝑟 𝑝 ), which contains p distinct solutions (configurations). Each configuration 𝑟 𝑗 = (𝜃 𝑗,1 , 𝜃 𝑗,2 ,...𝜃 𝑗,𝑖 ,...,𝜃 𝑗,𝑛+𝑚 ) is obtained randomly and uniformly (We use the generation of the initial population as in [31]). 𝜃 𝑗,𝑖 represents the ith variable decision of the jth configuration, where 𝜃 𝑗,𝑖 ( 𝑖 = 1,2, … , 𝑚 ) denote the buffer sizes and 𝜃 𝑗,𝑖 (𝑖 = 𝑚 + 1, … , 𝑛 + 𝑚) denote the service times.…”

“…In this study, the PR value is calculated by formulating perturbation generation and perturbation propagation rules which are intricately designed to anticipate the consequences of changes in buffer size in the NT, providing predictions, due to the introduction of perturbations, about the nature of future events and their respective durations. Earlier studies on serial manufacturing lines develop rules for generating and propagating perturbations in transfer/production cells [31], [35]. In this study, we provide an overview of these rules as applied to transfer cells (Fig.…”

Simultaneous allocation of buffer capacities and service times in unreliable production lines

“…Next, FPA iteratively replaces a current solution (found by GA) by a new one, until some stopping criterion is achieved. The detailed optimisation approach is presented in Figure 2 , where the initial population, , is composed of m different configurations (solutions), where each solution is generated uniformly and randomly (see Kassoul, Cheikhrouhou, and Zufferey ( 2022 ) for more details on the way to generate such a population of random solutions). represents the j th design parameter of the i th configuration, where are integer (resp.…”

“…Considering the simultaneous allocation of buffer capacities and service times, a contribution of this paper is to combine the Finite Perturbation Analysis (FPA) and the Genetic Algorithm (GA). Indeed, such methods have been proved to be efficient in the case of BAP (Kassoul, Cheikhrouhou, and Zufferey 2022 ). In line with other techniques having a learning mechanism (Schindl and Zufferey 2015 ; Thevenin and Zufferey 2019 ), GA has a good exploration ability as it can quickly generate a variety of solutions that are spread over a large portion of the solution space.…”

Production policy based on echelon base stocks for reverse logistics and industrial symbiosis: an IPA approach

“…The main advantage is then to be able to compute optimal values without the help of all information and without long computation times. When the two conditions of impartiality and convergence can no longer be verified, other methods of perturbation analysis must be developed [44] . Initially developed for queuing networks [43] , IPA has been applied for Markov processes.…”