Resequencing Problem Research Articles

Deep reinforcement learning for solving car resequencing with selectivity banks in automotive assembly shops

In the automobile manufacturing process, the complexity of the production environment and the different objectives of the upstream and downstream shops necessitate the resequencing of the car sequences using buffers. This paper addresses a car resequencing problem in automotive assembly shops by utilising selectivity banks to reshuffle the given upstream sequences to minimise the number of violations of sequencing rules in the downstream sequence. A new resequencing approach combining heuristics and deep reinforcement learning is proposed to solve the problem. Heuristics solve the automobile storage problem, while a deep reinforcement learning method based on proximal policy optimisation addresses the automobile release problem. Three state matrices, agent actions, and the reward function are designed based on the characteristics of the objective and selectivity banks. Additionally, a novel network structure combining convolutional neural networks and long short-term memory networks is proposed to train the agent. The results demonstrate that the proposed method outperforms two efficient deep reinforcement learning algorithms and two heuristics and obtains better solutions in handling dynamic events.

A deep Q network algorithm for a car resequencing problem in automobile factories

A deep Q network algorithm for a car resequencing problem in automobile factories

Integrating virtual resequencing with car resequencing via selectivity banks for mixed-model assembly lines

Physical resequencing and virtual resequencing are two general forms of sequence adjustment in the context of mixed-model assembly lines. Selectivity banks can be used to physically change the production sequences on a mixed-model assembly line. Virtual resequencing can be incorporated into this process in order to improve the reordering outcomes. This paper introduces a problem of car resequencing before final assembly via selectivity banks that incorporates virtual resequencing. A mathematical programming formulation is presented to describe this integrated resequencing problem. The objective is to minimize the number of violations of sequencing rules within the downstream sequence. Several heuristic approaches for the car storage and retrieval processes are proposed, based on which nine integrated resequencing algorithms are developed. Computational experiments show that incorporating virtual resequencing nearly always obtains noticeably better results than pure physical resequencing. In addition, one of the presented algorithms, denoted by IR9, which jointly applies a chromosome based fill procedure and an integrated release procedure, performs best.

Integrated real-time control of mixed-model assembly lines and their part feeding processes

Mixed-model assembly lines enable a flow-oriented mass production of individualized products such as cars. The most investigated operational decision problems in this domain are the sequencing problem, which plans the initial sequence of cars launched down the line, and the resequencing problem, which alters a given car sequence by applying resequencing buffers. This dichotomy of existing research neglects a third, even more short-term decision level: launch control under real-time conditions. In large automotive assembly lines unforeseen disturbances are rather the norm, so that one of the most important decisions to maintain a smooth and error-free production flow is the continuous launching decision of the next car model released from the central chassis storage system into final assembly. The launching decision in every cycle is made in a central control room (or leitstand) and has to consider the current state of the assembly line itself, with all potential time delays and defects in any assembly station, along with all part feeding processes under real-time conditions. This paper provides an integrated solution approach for real-time launch control that applies sophisticated optimization procedures considering both the workflow along the assembly line and its impact on part feeding processes. We test our real-time approach in a simulation study and compare it with alternative launching approaches. Our results show that sophisticated real-time optimization clearly outperforms its competitors. Furthermore, an integrated approach including part feeding aspects can considerably reduce safety stocks of parts within workstations.

Deep reinforcement learning for a color-batching resequencing problem

In automotive paint shops, changes of colors between consecutive production orders cause costs for cleaning the painting robots. It is a significant task to re-sequence orders and group orders with identical color as a color batch to minimize the color changeover costs. In this paper, a Color-batching Resequencing Problem (CRP) with mix bank buffer systems is considered. We propose a Color-Histogram (CH) model to describe the CRP as a Markov decision process and a Deep Q-Network (DQN) algorithm to solve the CRP integrated with the virtual car resequencing technique. The CH model significantly reduces the number of possible actions of the DQN agent, so that the DQN algorithm can be applied to the CRP at a practical scale. A DQN agent is trained in a deep reinforcement learning environment to minimize the costs of color changeovers for the CRP. Two experiments with different assumptions on the order attribute distributions and cost metrics were conducted and evaluated. Experimental results show that the proposed approach outperformed conventional algorithms under both conditions. The proposed agent can run in real time on a regular personal computer with a GPU. Hence, the proposed approach can be readily applied in the production control of automotive paint shops to resolve order-resequencing problems.

Semiglobal Sequence Alignment with Gaps Using GPU.

In this paper, we consider the pair-wise semiglobal sequence alignment problem with gaps, which is motivated by the re-sequencing problem that requires to assemble short reads sequences into a genome sequence by referring to a reference sequence. The problem has been studied before for single gap and bounded number of gaps. For single gap, there is a GPU-based algorithm proposed (Barton et al., 2015). In our work, we propose a GPU-based algorithm for the bounded number of gaps case, called GPUGapsMis. We implement the algorithm and compare the performance with the CPU-based algorithm, called CPUGapsMis. The algorithm has two distinct stages: the alignment phase, and the backtrack phase. We investigate several different approaches, in order to determine the most favorable for this problem, by means of a Hybrid model or a wholly-GPU based model, as well as the alignment of single text sequences or multiple text sequences on the GPU at a time. We show that the alignment phase of the algorithm is a good candidate for parallelization, with peak speedup of 11 times. We show that although the backtracking phase is sequential, it is more beneficial to perform it on the GPU, as opposed to returning to the CPU and performing there. When performing both phases on the GPU, GPUGapsMis achieves a peak speedup of 10.4 times against CPUGapsMis. Our data parallel GPU algorithm achieves results which are an improvement on those of an existing GPU data parallel implementation (Ojiaku, 2014).

Accelerated dynamic programming algorithms for a car resequencing problem in automotive paint shops

• Novel exact and heuristic algorithms for resequencing cars to minimize setup costs in automotive paint shops are proposed. • Incorporating strong lower and upper bounds in the dynamic programming framework reduces computation time significantly. • Proposed algorithms are more efficient than existing approaches and also more practically applicable. • Shows well-tailored exact algorithms can be more efficient than meta-heuristic algorithms. In this paper, a car resequencing problem (CRP) for automotive paint shops is considered, whereby a set of cars conveyed from an upstream shop to one of the multiple conveyors is retrieved sequentially before the painting operation. The aim of the CRP is to find a car retrieval sequence that minimizes the sequence-dependent changeover cost, which is the cost that is incurred when two consecutive cars do not share the same color. For this problem, we propose accelerated dynamic programming (ADP) algorithms that utilize strong combinatorial lower bounds and effective upper bounds in a standard dynamic programming framework, thus outperforming existing exact algorithms. Testing of our algorithms over a wide range of instances confirmed that they are more efficient than the existing approaches and are also more applicable in practice.

Low-Latency Dynamic Wavelength and Bandwidth Allocation Algorithm for NG-EPON

Recently, the IEEE 802.3ca Task Force has been working to standardize the next-generation Ethernet passive optical network (NG-EPON) to fulfill future bandwidth demands. NG-EPON allows an optical network unit (ONU) transmission to be scheduled on multiple wavelength channels simultaneously. This will increase network capacity up to 100 Gb/s, along with flexibility, as any of the channels can be allocated from the four wavelength channels, each having a capacity of 25 Gb/s. However, it also raises many technical aspects to be considered while deciding on an optimum scheduling strategy for NG-EPON (100G-EPON) architecture. This paper proposes a dynamic wavelength and bandwidth allocation (DWBA) algorithm for NG-EPON to schedule upstream wavelength channels while providing a low average packet delay and catering frame resequencing problem. An online scheduling framework with gated service grant sizing policy has been demonstrated in this work. The proposed DWBA algorithm’s performance is evaluated in comparison with legacy interleaved polling with an adaptive cycle time (modified- IPACT) and recently proposed water-filling algorithmbased DWBA (WF-DBA) algorithms. Simulation results confirmed the validity of the proposed DWBA algorithm.

Resequencing with parallel queues to minimize the maximum number of items in the overflow area

AbstractThis article treats an elementary optimization problem, where an inbound stream of successive items is to be resequenced with the help of multiple parallel queues in order to restore an intended target sequence. Whenever early items block the one item to be currently released into the target sequence, they are withdrawn from their queue and intermediately stored in an overflow area until their actual release is reached. We aim to minimize the maximum number of items simultaneously stored in the overflow area during the complete resequencing process. We met this problem in industry practice at a large German automobile producer, who has to resequence containers with car seats prior to the assembly process. We formalize the resulting resequencing problem and provide suited exact and heuristic solution algorithms. In our computational study, we also address managerial aspects such as how to properly avoid the negative effects of sequence alterations. © 2016 Wiley Periodicals, Inc. Naval Research Logistics 63: 401–415, 2016

Paint batching problem on M-to-1 conveyor systems

An M-to-1 conveyor system consists of multiple upstream conveyors and a single downstream conveyor. In this paper, we investigate the paint batching problem on M-to-1 conveyor systems with the objective of minimizing setup costs. Our research is motivated by a vehicle re-sequencing problem at a major Korean automotive manufacturer. Setup costs are incurred when two consecutive jobs in the downstream conveyor do not share the same feature. Re-sequencing flexibility is limited by the precedence relationship among jobs in the upstream conveyors. First, we develop a mixed integer linear programming model and propose an efficient dynamic programming (DP) algorithm for a 2-to-1 conveyor system. However, because the suggested DP cannot guarantee optimality in general settings, we propose two efficient genetic algorithms (GAs) to find near optimal solutions. Specifically, we design the reordering operation for making offspring to satisfy the precedence condition. We show that the proposed GAs perform prominently with respect to optimality gap and computation time; thus, they are amenable to environments where solutions must be obtained within tight time constraints.

Dynamic resequencing at mixed-model assembly lines

In this paper, we investigate a dynamic resequencing problem covering realistic properties of a mixed-model assembly line. To this end, we present a mathematical model that addresses dynamically supplied blocking information and viable due dates. We developed two different strategies that use a static resequencing algorithm as a subroutine. One strategy integrates each unblocked order immediately into the planned sequence, whereas the other strategy waits for good positions that do not conflict with the due dates. All algorithms construct guaranteed feasible sequences. Using industrial test data, we show that both strategies perform significantly better than a simple method derived from practice. A replanning procedure that tries to improve the current planned sequence whenever computing time suffices yields an additional improvement for both strategies.

Solving the Car Resequencing Problem with mix banks

Mixed-model assembly lines are used at the automotive industry to obtain more efficient production flows. The sequencing of the different units is done at the very beginning of the process, but frequently large perturbations of the original sequence can be observed. These perturbations might be intended or not but to some extent, a resequencing process –to create a new sequence with the available units- is necessary. This paper focuses in the re-sequencing process by using parallel buffering lines (known as mix banks). Algorithms to decide the line of destination of incoming units and the line of extraction of units to the line are presented and evaluated. The algorithms and their performance have been verified using discrete event simulation software (SIMIO). An analysis considering different operations situations is done and some recommendations to real case users are presented.Keywords: Mixed-model assembly line, Car sequencing, Resequencing, Mix bank, Dynamic algorithm.Resolviendo el problema de Resecuenciación de Coches con el uso de líneas de acumulo paralelasResumen: En la industria del automóvil se utilizan líneas mixtas de montaje de modelos para obtener flujos de producción más eficientes. Aunque la secuencia de los productos en la línea se contempla en el comienzo del proceso, se suelen producir perturbaciones en la secuencia de producción. En ocasiones es necesaria una resecuenciación entendida como la creación de una nueva secuencia. Este trabajo se centra en la resecuenciación mediante el uso de líneas paralelas de almacenamiento (conocidas como mix-bank). Se proponen algoritmos para decidir en qué línea se debe introducir y de qué línea se debe extraer. Los algoritmos son verificados mediante modelos de simulación implementados en SIMIO. Se realiza un análisis para diferentes factores operacionales y se proponen reglas de gestión para su uso en casos reales.Palabras Claves: Líneas mixtas de montaje, Secuenciación de Coches, Resecuenciación, Mix Bank, Algoritmo dinámico.

A decomposition approach for the car resequencing problem with selectivity banks

An important decision problem when mass-producing customized product to order is the sequencing problem, which decides on the succession of models launched down a mixed-model assembly line. To avoid work overload of workforce the car sequencing problem restricts the maximum occurrence of labor-intensive options, e.g., a sunroof, in a subsequence of a certain length by applying sequencing rules. In the real-world, frequently perturbations occur stirring up an initially planed sequence, so that a resequencing is required. This paper treats the car resequencing problem where a selectivity bank, which is a special form of buffer organization consisting of parallel line segments without assembly operations, is applied to reshuffle a given initial sequence and rule violations are to be minimized. The problem is formalized and suited heuristic solution procedures are presented and tested. Furthermore, the impact of differently sized mix-banks on resequencing flexibility is investigated.

Limited Re-Sequencing for Mixed-Models with Multiple Objectives, Part II: A Permutation Approach

This research presents an approach to solving the limited re-sequencing problem for a JIT system when two objectives are considered for multiple processes. One objective is to minimize the number of setups; the other is to minimize the material usage rate [1]. For this research effort, each unique permutation of the problem’s demand structure is noted, and used as a mechanism for finding subsequent sequences. Two variants of this permutation approach are used: one employs a Monte-Carlo simulation, while the other employs a modification of Ant-Colony Optimization to find sequences satisfying the objectives of interest. Problem sets from the literature are used for assessment, and experimentation shows that the methodology presented here outperforms methodology from an earlier research effort [3].

The Car Resequencing Problem with Pull-Off Tables

AbstractThe car sequencing problem determines sequences of different car models launched down a mixed- model assembly line. To avoid work overloads of workforce, car sequencing restricts the maximum occurrence of labor-intensive options, e.g., a sunroof, by applying sequencing rules. We consider this problem in a resequencing context, where a given number of buffers (denoted as pull-off tables) is available for rearranging a stirred sequence. The problem is formalized and suited solution procedures are developed. A lower bound and a dominance rule are introduced which both reduce the running time of our graph approach. Finally, a real-world resequencing setting is investigated.

Level scheduling under limited resequencing flexibility

A mixed-model assembly line requires the solution of a short-term sequencing problem, which decides on the succession of different models launched down the line. A famous solution approach stemming from the Toyota Production System is the so-called Level Scheduling (LS), which aims to distribute the part consumption induced by a model sequence evenly over the planning horizon. LS attracted a multitude of different researchers, who, however, invariably treat initial sequence planning where all degrees of freedom in assigning models to production cycles exist. In the real-world, conflicting objectives and restrictions of preceding production stages, i.e., body and paint shop, simultaneously need to be considered and perturbations of an initial sequence will regularly occur, so that the sequencing problem often becomes a resequencing problem. Here, a given model sequence is to be reshuffled with the help of resequencing buffers (denoted as pull-off tables). This paper shows how to adapt famous solution approaches for alternative LS problems, namely the Product-Rate-Variation (PRV) and the Output-Rate-Variation (ORV) problem, if the (re-)assignment of models to cycles is restricted by the given number of pull-off tables. Furthermore, the effect of increasing re-sequencing flexibility is investigated, so that the practitioner receives decision support for buffer dimensioning, and the ability of the PRV in reasonably approximating the more detailed ORV in a resequencing environment is tested.

Sequencing with limited flexibility

We consider the problem of resequencing a set of prearranged jobs when there is limited resequencing flexibility and sequence-dependent changeover costs. Resequencing flexibility is limited by how far forward or backward a job can shift in the sequence relative to its original position. We show how the problem can be solved using dynamic programming in polynomial time with respect to the number of jobs. We also show how the same solution approach can be extended to problems where sequencing constraints are job specific and to problems where job features, which determine changeover costs, are jointly determined with the job sequence. We provide an integer programming formulation to the resequencing problem whose linear programming relaxation offers a useful lower bound. We also describe a family of decomposition heuristics that are easy to customize to provide desired levels of solution quality and solution time. We document the quality of the lower bound from the linear programming relaxation and the upper bound from the heuristic using numerical results. We also provide numerical results to support managerial insights regarding the value of flexibility. We show that the value of flexibility is of the diminishing kind with most of the benefit realized with relatively limited flexibility. We also show that a balanced allocation of flexibility among forward and backward position shifting is superior to an unbalanced one. More significantly, we show that forward and backward position shifting flexibility are complements with the value of one increasing in the amount of the other. Finally, we apply our solution approach to a real-world case from the automotive industry.

Improvement of earliness and lateness by postponement on an automotive production line

This article takes an interest in improving the performance of a subproblem of car sequencing: the resequencing problem. It extends a model of assembly line with the simulation of buffer stocks with drawers. Presequencing and resequencing are modeled with different types of stock management policy. Furthermore, postponement is introduced by two points of differentiation within the line. After choosing relevant variables and performance indicators a large-scale design of experiments finally enables to estimate average improvement brought by postponement to more than 39.7%, compared to the situation without postponement.

Permutations generated by token passing in graphs

A transportation graph is a directed graph with a designated input node and a designated output node. Initially, the input node contains an ordered set of tokens 1,2,3, … The tokens are removed from the input node in this order and transferred through the graph to the output node in a series of moves; each move transfers a token from a node to an adjacent node. Two or more tokens cannot reside on an internal node simultaneously. When the tokens arrive at the output node they will appear in a permutation of their original order. The main result is a description of the possible arrival permutations in terms of regular sets. This description allows the number of arrival permutations of each length to be computed. The theory is then applied to packet-switching networks and has implications for the resequencing problem. It is also applied to some complex data structures and extends previously known results to the case that the data structures are of bounded capacity. A by-product of this investigation is a new proof that permutations which avoid the pattern 321 are in one to one correspondence with those that avoid 312.

Analysis of the resequencing buffer in a homogeneous M/ M/2 queue

The resequencing problem is encountered in many practical information systems such as distributed database and communication networks. In these systems customers, such as messages in a computer network, have to be delivered to users in their original order. Therefore, those customers which become out of order due to the randomness of the system are forced to wait in a resequencing buffer so that their delivered order can be guaranteed. The previous work on the resequencing problem mainly concentrated on the delay aspect. From both theoretical and practical viewpoints, however, the queue length characteristics of the resequencing buffer are also significant. We consider the queue length distribution of the resequencing buffer fed by a homogeneous M/ M/2 queue. The exact analysis is carried out for the probability mass functions of the queue length in equilibrium and the maximal occupancy which corresponds to the queue length just before the departure instants of customers from the resequencing buffer.