Search

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2007.02.055 Byline: Tiente Hsu (a), Ouajdi Korbaa (b), Remy Dupas (a), Gilles Goncalves (a) Keywords: Flexible manufacturing systems; Cyclic scheduling; Genetic algorithm; Petri nets Abstract: This paper concerns the domain of flexible manufacturing systems (FMS) and focuses on the scheduling problems encountered in these systems. We have chosen the cyclic behaviour to study this problem, to reduce its complexity. This cyclic scheduling problem, whose complexity is NP-hard in the general case, aims to minimise the work in process (WIP) to satisfy economic constraints. We first recall and discuss the best known cyclic scheduling heuristics. Then, we present a two-step resolution approach. In the first step, a performance analysis is carried out; it is based on the Petri net modelling of the production process. This analysis resolves some indeterminism due to the system's flexibility and allows a lower bound of the WIP to be obtained. In the second step, after a formal model of the scheduling problem has been given, we describe a genetic algorithm approach to find a schedule which can reach the optimal production speed while minimizing the WIP. Finally, our genetic approach is validated and compared with known heuristics on a set of test problems. Author Affiliation: (a) Laboratoire de Genie Informatique et Automatique de l'Artois (L.G.I.2.A), Universite d'Artois Faculte des Sciences Appliquees, Technoparc Zone Futura BP 819, 62400 Bethune, France (b) Institut Superieur d'Informatique et des Technologies de Communication de Hammam Sousse, 5 bis rue 01 Juin 1955, 4011 Hammam Sousse, Tunisia Article History: Received 27 January 2004; Accepted 1 February 2007

In this note, we consider a fluid flow, single part-type, single unreliable machine production system with a bounded backlog/inventory space. We prove that, as in the unbounded case, the problem of minimizing an infinite horizon average demand loss/backlog/surplus cost is solved by a hedging point policy. An implicit equation is given, whose structure easily allows to numerically evaluate the optimal safety stock. The effect of system parameters on the optimal safety stock is analyzed and some numerical examples illustrate the presented results. Index Terms--Bounded inventory, hedging point, limited backlog, unreliable machine.

Bounds on the size of the plant uncertainties are found such that the use of the inversion-based feedforward input improves the output-tracking performance when compared to the use of feedback alone (i.e., without the feedforward). The output-tracking error is normalized by the size of the desired output and used as a measure of the output tracking performance. The worst-case performance is compared for two cases: 1) with the use of feedback alone and 2) with the addition of the feedforward input. It is shown that inversion-based feedforward controllers can lead to performance improvements at frequencies [omega] where the uncertainty [DELTA](j[omega]) in the nominal plant is smaller than the size of the nominal plant [G.sub.0](j[omega]) divided by its condition number [[kappa].sub.[G.sub.0]](j[omega]), i.e., [parallel] [DELTA](j[omega][[parallel].sub.2] < [parallel][G.sub.0](j[omega]) [[parallel].sub.2]/[[kappa].sub.[G.sub.0]](j[omega]). A modified feedforward input is proposed that only uses the model information in frequency regions where plant uncertainty is sufficiently small. The use of this modified inverse with (any) feedback results in improvement of the output tracking performance, when compared to the use of the feedback alone. Index Terms--Feedforward, nonminimum phase, output-tracking, robust, system-inverse.

Asynchronous production systems are classified into systems with fixed, unrestricted, and restricted work-in-process inventory. For each class routing optimization models are presented. Furthermore, extensions are provided in order to include optimal lotsizing and/or part mix decisions. These extensions find optimal solutions for two previously published problems with only heuristic solution procedures., 1. Introduction This paper analyses models for optimal order release decisions in asynchronous production systems. The latter are production systems which are able to process a variety of part types. [...]

Shape deposition manufacturing (SDM) is a solid freeform fabrication (SFF) methodology for automatically building up material layers to form three-dimensional, complex-shaped, multi-material structures. Microcasting is a molten metal droplet deposition process which is able to create fully dense metal layers with controlled microstructures. SDM combines microcasting with other intermediate processing operations, such as CNC machining and shot peening, to create high quality metal parts. In this paper, a description is given of SDM and the microcasting process. An overview of thermal and mechanical issues associated with SDM and microcasting is presented, including the control of interlayer metallurgical bonding through substrate remelting, the control of cooling rates of both the substrate and the deposited material and the minimization of residual thermal stress effects. Thermal models are used to study the issue of localized remelting of previously deposited material by newly deposited molten droplets to achieve metallurgical bonding. Mechanical modeling provides insight into residual stress build-up during part manufacture and residual stress-driven debonding between deposited layers.

This paper presents some results of an experimental study on chip breaking monitoring when turning steels with chip forming inserts. In the first part of this study the recognition of chip forths by means of cutting force analysis in the frequency domain is described. Generally, the present work is concerned with the use of the interface control factor for determining the transition from unacceptable chip forms to the chip breaking. For this purpose experiments under approximately orthogonal machining conditions including measuring the shear strain, the cutting energy and the average interface temperature are carried out. The obtained results confirmed that the initiation of acceptable chip breaking for a wide range of cutting conditions and the three steels used could be predicted in terms of an appropriate energy consumption rate. An architecture for on-line cutting state monitoring system based on signal processing procedure in the frequency domain is developed and the performance of the system is experimentally evaluated.

This paper presents a mathematical model for predicting the natural resting aspect of parts of regular shapes but of different sizes based on stability considerations. The objective of this model is to facilitate the design of effective and efficient orientating devices for vibratory bowl feeders. The results from the mathematical modelling were compared with the theoretical models and experimental results of other investigators. Good agreement was obtained.

This paper describes a pretravel model for vertically-oriented touch trigger probes with straight styli widely used on coordinate measuring machines (CMMs). Advances in machine accuracy and tighter part tolerances leave probe error as one of the major error sources in dimensional measurement processes. Pretravel accounts for the majority of touch trigger probe errors and is caused by bending deflection of the probe stylus shaft. A trigger force model was first derived and used to model the probing force at the trigger instant. Experimental probing force data were used to validate the trigger force model. The trigger force model was then used to model the bending deflection of the stylus shaft at the trigger instant. Experimental x-y-z coordinate data were used to validate the pretravel model. It is shown that 24 points of equatorial data measured in the probe calibration process can be used to accurately identify the only model parameter. Also, 12 points of equatorial data provides reasonably good results. Model predicted pretravels associated with 1561 probe approach directions were compared with experimental pretravels (range: 16 [[micro]meter], when using a 50 mm stylus) and the standard deviation of prediction errors can be reduced to around 0.6 [[micro]meter].

This article presents a method for characterizing geometric shape variations in manufactured automotive spaceframe extrusions using measurement data obtained by a Coordinate Measurement Machine (CMM). With a manufacturing perspective, a modeling scheme is proposed for representing the actual shape of the extrusions. Systematic deviations in the geometry of the parts from their theoretically perfect design description models are described by the critical parameters of their manufactured part models (MPMs), which are constructed by fitting CMM data through an optimization. The manufactured part model can then be interpreted in a meaningful way for improving bending process as well as quantifying the process capability. An application of the methodology is illustrated through the analysis of experimental data for rotary draw bent extrusions used in the aluminum intensive automobiles.

Autocorrelation in 100 percent measurement data results in false alarms when the traditional control charts, such as X and R charts, are applied in process monitoring. A popular approach proposed in the literature is based on prediction error analysis (PEA), i.e., using time series models to remove the autocorrelation, and then applying the control charts to the residuals, or prediction errors. This paper uses a step-function type mean shift as an example to investigate the effect of prediction error analysis on the speed of mean shift detection. The use of PEA results in two changes in the 100 percent measurement data: (1) change in the variance, and (2) change in the magnitude of the mean shift. Both changes affect the speed of mean shift detection. These effects are model parameter dependent and are obtained quantitatively for AR (1) and ARMA (2,1) models. Simulations and examples from automobile body assembly processes are used to demonstrate these effects. It is shown that depending on the parameters of the AMRA models, the speed of detection could be increased or decreased significantly.

A general kinematically admissible velocity field is suggested for the upper-bound solution of hydroforming of arbitrarily-shaped boxes. The suggested formulation is then applied to hydroforming of an elliptic-circular box. From the proposed velocity field, the fluid pressure vs. punch stroke relationship to render uniform thickness and the deformed configuration are determined by minimizing the total power consumption with respect to some chosen parameters. Experiments are carried out in the hydroforming press according to the computed pressure vs. punch stroke curve. The assumption of uniform wall thickness is confirmed by measuring the thickness variation. The effects of various process parameters including blank size, work-hardening exponent and frictional coefficient on the pressure curve are analyzed and discussed. It is thus shown that the proposed method of analysis in the present study can be effectively used for hydroforming of arbitrarily-shaped boxes.

In this paper, we present a method for the selection of processes to manufacture various parts of an assembly by establishing a compromise between product quality and part manufacturing cost. We quantify the impact the precision of a part characteristic has on the overall quality of a product by using a standard Taguchi loss function. Part manufacturing cost is modeled as a function of process precision (i.e., standard deviation of the output characteristic) as opposed to previous models where manufacturing cost is a function of part tolerance. This approach is more realistic and does not assume, a priori, a relationship between conventional tolerance and process spread. Rather than allocating conventional tolerances on the assembly parts, we use statistical tolerances that are more pertinent when using a quality loss function. The model adopted makes it possible to investigate the relationship between optimum quality loss and tolerance variations. As expected, the optimum quality loss generally decreases when the tolerance increases. Exceptions may be encountered when changes of process occur. The manufacture of a simple three component assembly is studied to illustrate the findings.

An innovative methodology for the essential supposition of product design was developed based on the feedback loop. The feedback approach was utilized to evaluate and select design variants over all stages of product design according to product requirements. The model that was developed based on the methodology can also be utilized for cost estimations involving sophisticated and complex structures. Furthermore, this determination can be utilized for decision support and early estimations of future product costs.

The adoption of Japanese-influenced production techniques by US manufacturers has drastically improved the performance of American industries. One of these methods include the just-in-time (JIT) system developed and perfected by Toyota. A model that applies full and partial JIT application show major improvements in plant performance, such as downtime, cycle time, lead time, percentage passed and percentage shipped.

The concept of more maintenance effort (or training) implying higher system reliability (or efficiency) is well accepted. In this paper we study an optimal maintenance (training) program based on this concept for a simple production system using the M/G/1 vacation queue model. We discuss how to determine the optimal maintenance (training) time to minimize the average time spent by a job in the system., 1. Introduction Consider a production system where jobs arrive for processing by a machine according to a certain stochastic process, and the machine is subject to random breakdowns and repairs. [...]

We extend recent works by Tang and Tang dealing with the quality cost of a process that is screened with the use of correlated variables because the principal quality characteristic is difficult or expensive to measure. Our model includes a very flexible and general set of loss functions that can be used to model both traditional and contemporary 'losses' for deviation from target values or product specifications. We also give an analysis of the expected value of perfect information and sample information obtained from the screening variables. This facilitates the evaluation of the advisability of using screening variables., 1. Introduction There are quality characteristics that are expensive, destructive, or impossible to measure in the finished product. There may, however, be variables correlated to these quality characteristics that can [...]

Among the many changes required in implementing a just-in-time (JIT) production system, the change from batch to level production was the most difficult part in a consumer electronics company in Korea. Simulation was used to persuade workers of the importance of level production by comparing the performance of batch production to level production. This article describes how simulation was used to convince workers to change and later to show them the improvements resulting from that change. KEYWORDS: discrete simulation; JIT; level production.

In assembly line manufacturing systems and material handling systems the operation is usually done by a system of serially linked stages. It is apparent that in a serially dependent system of this type the failure of any stage either causes a reduction or causes the complete shut down of the operation of the system. The provision of buffers between stages can result in operation of the nonfailed stages even in the event of failure of a stage. The use of buffers therefore increases the system availability. The availability of buffered systems is a function of the size of the buffer, and it is maximized with a buffer of size infinity. However, because of the infeasibility of having infinite buffers the maximum possible availability is rarely attained. This paper presents a stochastic analysis of a buffered production system. The probability of the buffer being full is determined as a function of buffer capacity. This in turn has been used to make the decision on buffer size. The analysis in this paper indicates that, over the range of buffer capacity from zero to infinity, the probability of the buffer-full-state exhibits monotonically decreasing behavior that asymptotes to zero as buffer size approaches infinity. This finding can be used to detect the range over which the probability of the buffer-full-state asymptotically approaches zero, and appropriately select a buffer size beyond which no significant gains in system availability is possible to justify the incremental cost of the additional buffer capacity.

A solution to the lotsizing problem insingle bottleneck systems is developed. Such systems are multistage production systems wherein only one setup is nonzero. The bottleneck, which refers to the stage with the nonzero setup, is a critical resource because its throughput determines the throughput of the entire system. Goldratt and Cox (1986) and Lundrigan (1986) proposed the OPT approach to this problem, in which productionproceeds with binomial yield lotsizes. Multiple production runs are required given that full satisfaction of demand is required. Assuming binomial yields, the optimal control problem is reducible to single-stage optimal lotsizing. However, solution to the problem versions that assume more than one bottleneck are yet to be examined.

Optimal flows for flexible manufacturing cells with pipeline operations are modeled. Resource allocation and operation scheduling are examined in the context of the cyclic production of a product with a given series of production operations and a specified set of resource requirements. Ageneral model with special applications to polynomial or NP-complete problems is proposed. The numbers of both operating machines and product units that are operational at each time period are varied in order to generate different cases. The polynomial solution exists given two streams when independent assignment of all operations to the robots is possible, thereby leading to a result expressed in terms of the shortest path on a specific network.