Multi-product Manufacturing Research Articles

Applying Bond Graphs for Modelling the Manufacturing Dynamics

Supply chain dynamics and execution control seem not to receive as much attention as the domain of supply chain planning by means of static optimization models. The dynamics of manufacturing systems also deserves attention, since they are the fundamental pieces of a supply chain. In this paper, a dynamic model for the production control of a multi-product manufacturing system is developed using the Bond Graph modeling methodology. The proposed model is applied to depict the dynamics of a real job shop production system of propylene bags, where the control objective is to adjust the processing frequency of the machines to attend the demands of the products, while keeping the work in process at the desired levels. Exploratory simulations were carried out, and the preliminary results showed that the system could be autonomously controlled. Some advantages of the Bond Graph methodology are also highlighted.

Decentralized Manufacturing Supply Chains Coordination under Uncertain Competitiveness

The coordination of decentralized multi-product manufacturing SCs is achieved through negotiations based on expected win-win principles in an uncertain competitive environment. Based on non-symmetrical roles of the different actors, the client (as leader) is supposed to propose coordination contracts according to its best expected conditions, taking into account the uncertain reaction of the provider (follower). This uncertain reaction is modeled as a probability of acceptance, computed according to the overall scenario conditions, which include the presence of 3rd parties. Different negotiation scenarios are analyzed considering cooperative and non-cooperative cases. The resulting MINLP tactical models are illustrated using a case study with different providers (follower SC) around a client (leader SC) interacting in a global decentralized scenario. The negotiations based on non-cooperative cases proves to identify the situation with higher independent profit expectations. Moreover, the proposed approach shows the importance of considering the uncertainty associated with the response of the follower to the leader's decisions, resulting from a wider knowledge of its options.

Industrial process system assessment: bridging process engineering and life cycle assessment through multiscale modeling

Conducting life cycle assessments (LCA) of complex multi-line, multi-product manufacturing facilities can be challenging. There is no standard method to assign or allocate resource use and environmental impacts to the individual products. The Industrial Process System Assessment method is presented to address this gap in current approaches for LCA allocation by making use of a multiscale model of process systems. Algorithms for the allocation procedure rectify process, ancillary, and non-process data from a facility such that inputs (e.g., raw materials) and outputs (e.g., emissions) can be assigned to a particular product from a particular line. The allocation approach demonstrates systematic multiscale process model development to avoid arbitrary allocation assumptions. This standardized method will enable LCA practitioners and process engineers to more clearly differentiate resource use and environmental impacts of co-products from complex facilities. The model could be further extended to provide product allocation based on individual unit operations when data are available.

Optimal supplier selection and order allocation for multi-product manufacturing featuring customer flexibility

Efficient and effective supplier-selection and order-allocation decisions are critical for manufacturing industries to ensure stable material flows in today’s highly competitive supply chain, in particular, when customers are willing to accept products with less desirable attributes (e.g., colour, material) for economic reasons. This study terms this kind of customer behaviour as ‘customer flexibility’ and attempts to optimally solve the challenging problem of supplier selection and order allocation incorporating customer flexibility. A new mixed integer programming model is developed to maximise manufacturer’s total profit. Due to the complexity and non-deterministic polynomial-time -hard nature of the problem, a novel hybrid constraint programming (CP) and simulated annealing (SA) algorithm ‘CP–SA’ is developed to solve the problem optimally. Extensive computational experiments clearly demonstrate its excellent performance.

Analysis of multi-product manufacturing systems with arbitrary processing times

Multi-product systems with finite buffers and sequence-dependent set-up times are quite common in modern manufacturing industry. In practice, the distribution of machine processing time could be arbitrary, while in existing literature it is often assumed to follow an exponential distribution. In this paper, we develop an analytical method to study the multi-product manufacturing systems with non-exponential processing times. An embedded Markov chain model is constructed and two approximation methods, Gamma estimation and linear approximation, are proposed. The model is validated with high accuracy by numerical experiments and practical data from an automotive assembly system.

Capacity optimization and scheduling of a multiproduct manufacturing facility for biotech products.

A general mathematical framework has been proposed in this work for scheduling of a multiproduct and multipurpose facility involving manufacturing of biotech products. The specific problem involves several batch operations occurring in multiple units involving fixed processing time, unlimited storage policy, transition times, shared units, and deterministic and fixed data in the given time horizon. The different batch operations are modeled using state-task network representation. Two different mathematical formulations are proposed based on discrete- and continuous-time representations leading to a mixed-integer linear programming model which is solved using General Algebraic Modeling System software. A case study based on a real facility is presented to illustrate the potential and applicability of the proposed models. The continuous-time model required less number of events and has a smaller problem size compared to the discrete-time model.

Analysis of Multiproduct Manufacturing Systems With Homogeneous Exponential Machines

This paper is devoted to modeling and analysis of multiproduct manufacturing systems with homogeneous exponential machines and finite buffers. In such systems, each machine processes multiple product types with different speeds, but the processing time for each product type is the same on all machines. Buffers are finite, shared for all products. Analytical methods to evaluate the system performance are developed, and system-theoretic properties are investigated. It is shown from numerical experiments that such a method has a high accuracy in performance evaluation. To improve the performance of such systems, bottleneck analysis is carried out to identify the machine and/or product whose improvement will lead to the largest improvement in system throughput.

Оптимизация технологических операций сборки в роботизированном комплексе ГАСК электронных модулей из разнотипных компонентов многономенклатурного производства

Оптимизация технологических операций сборки в роботизированном комплексе ГАСК электронных модулей из разнотипных компонентов многономенклатурного производства

Проектирование и оптимизация технологических структур роботизированного комплекса ГАСК сборки электронных модулей в многономенклатурном производстве

Проектирование и оптимизация технологических структур роботизированного комплекса ГАСК сборки электронных модулей в многономенклатурном производстве

Automated guided vehicles fleet match-up scheduling with production flow constraints

The paper describes a Multimodal Transportation Network (MTN) in which several unimodal networks (Automated Guided Vehicles (AGVs), hoists, lifts, etc.) interact with each other via common shared workstations as to provide a variety of demand-responsive material handling operations. The material handling transport modes provide movement of work pieces between workstations along their manufacturing routes in the MTN. The main contribution of this work is the solution to a constraint satisfaction problem aimed at AGVs fleet match-up scheduling while taking into consideration assumed itineraries of concurrently manufactured product types. In other words, assuming a given topology of the MTN and schedules of operation sequences modeling concurrently manufactured product types, the main objective is to provide a declarative framework aimed at determining conditions allowing one to adjust the AGVs fleet schedule due to the timetable of operations executed in an assumed multi-product manufacturing environment.

Best practices for successful implementation of single‐use technology and qualification

Single‐use bioprocessing facilities are providing flexible capacity for clinical trials and low‐ to mid‐volume market supply. However, while this production technology reduces manufacturing cost, quantum jumps in healthcare cost efficiency can be expected from new product technology. In the future, new product technology such as virus‐based systems will need smaller bioprocessing capacities which may be provided by single‐use facilities. The single‐use biomanufacturing facility at Rentschler Biotechnologie received the 2012 Facility of the Year Award in the category Equipment Innovation. The award recognizes the project for its flexible, multiproduct manufacturing facility designed to minimize manufacturing costs and product cycle times. This clinical manufacturing facility proves concepts that pave the way toward the factory of the future for commercial biologics manufacturing. From that and more recent projects, we derived the best practices when implementing such a facility during all project phases (when assessing its feasibility, designing, constructing, qualifying, starting up, and operating it). Challenges, potential solutions, and future single‐use bioprocessing facility design objectives are discussed.

Comparison between production controls in multi-stage multi-product manufacturing environments: two case studies

Objective: The paper presents a study on parallel CONWIP systems based on two case studies benchmarked against CONWIP and push. The performance measures are flow time and work-in-process (WIP) level.Methodology: A discrete event simulation model was constructed and was used to run experiments for individual production controls. The gathered data were analyzed through response surface methodology.Results: In case study 1, the parallel CONWIP with FIFO dispatch rule is superior at both flow time and WIP level. In case study 2, push, CONWIP, and parallel CONWIP with LH dispatch rule to be considered superior at both flow time and WIP level.Conclusion: One notable advantage of parallel CONWIP systems is the ability of the card count to respond to changes in the demand for product classes. Several future research directions are proposed, such as framework of dynamic card controlling, and implementation of parallel CONWIP aligned with lean practices.

A comparison of make-to-stock and make-to-order in multi-product manufacturing systems with variable due dates

This article models a single-stage hybrid production system, which can be regarded as a Make To Order (MTO) production system with safety stocks or a Make To Stock (MTS) production system with advance demand information. In an environment with multiple products and variable customer due dates, optimality conditions for safety stocks (base stocks) and safety lead times (work-ahead window) that minimize inventory and backorder costs are derived. For a simplified M/M/1 system with exponentially distributed customer required lead time, an explicit comparison between MTO and MTS is conducted. A pure MTO policy gets relatively more favorable to a pure MTS policy if inventory holding costs increase, backorder costs decrease, the mean customer required lead time increases, or the processing rate increases. In a numerical study, the influence of variance, the behavior of optimal parameters, and the cost reduction potential of this hybrid policy are shown.

Integrated Product Blending Optimization for Oil Refinery Operations

Refinery is a multiproduct manufacturing plant. Crude oil is processed to get intermediate products which are blended to meet quantity, quality and schedule specification of the final products. Refinery optimization is a complex problem therefore it is broken down into sub-problems which are solved independently. The solution obtained using this approach can be improved by integrating different sub-problems for real-time optimization. Refinery works on very small Gross Refinery Margin (GRM); off specification product blends results in considerable cost overhead because product blending is the last operation in the refinery process. Therefore we propose a method for real time integration of product blending with secondary process units to significantly improve the GRM. This method is designed and implemented in the present work. A case study is included to demonstrate the benefits of the proposed method.

Step-by-step cyclic processes scheduling

Automated Guided Vehicles (AGVs) fleet scheduling is one of the big problems in Flexible Manufacturing System (FMS) control. The problem is more complicated when concurrent multi-product manufacturing and resource deadlock avoidance policies are considered. The objective of the research is to provide a declarative model enabling to state a constraint satisfaction problem aimed at AGVs fleet scheduling subject to assumed itineraries of concurrently manufactured product types. In other words, assuming a given layout of FMS's material handling and production routes of simultaneously manufactured orders, the main objective is to provide the declarative framework aimed at conditions allowing one to calculate the AGVs fleet schedule in online mode. An illustrative example of the relevant algebra-like driven step-by-stem cyclic scheduling is provided.

A Bernoulli Model of Multi-Product Manufacturing Systems

Modeling and analysis of multiple products manufacturing systems are of significant importance. In this paper, a Bernoulli model of multi-product manufacturing system with unreliable machines and finite buffers is introduced. In such a model, each machine is capable of processing multiple product types, and each buffer is shared for all types of products. All machines process the parts following the same sequence of different product types. By aggregating different product types into a single one, closed formulas have been derived to evaluate the production rate of the line with one or two machines, and recursive procedures are used to analyze longer lines. Numerical studies indicate that such a method has a high precision in performance evaluation. In addition, system-theoretic properties, such as asymptotic property, monotonicity and reversibility, have been investigated.

Scheduling policies in multi-product manufacturing systems with sequence-dependent setup times and finite buffers

Multi-product production systems with sequence-dependent setup times are typical in the manufacturing of semiconductor chips and other electronic products. In such systems, the scheduling policies coordinating the production of multiple product types play an important role. In this paper, we study a multi-product manufacturing system with finite buffers, sequence-dependent setup times and various scheduling policies. Using continuous-time Markov chain models, we evaluate the performance of such systems under seven scheduling policies, i.e. cyclic, shortest queue, shortest processing time, shortest overall time (including setup time and processing time), longest queue, longest processing time, and longest overall time. The impacts of these policies on system throughput are compared, and the conditions characterising the superiority of each policy are investigated. The results of this work can provide production engineers and supervisors practical guidance to operate multi-product manufacturing systems with sequence-dependent setups.

The robustness of scheduling policies in multi-product manufacturing systems with sequence-dependent setup times and finite buffers

In this paper, a continuous time Markov chain model is introduced to study multi-product manufacturing systems with sequence-dependent setup times and finite buffers under seven scheduling policies, i.e., cyclic, shortest queue, shortest processing time, shortest overall time (including setup time and processing times), longest queue, longest processing time, and longest overall time. In manufacturing environments, optimal solution may not be applicable due to uncertainty and variation in system parameters. Therefore, in this paper, in addition to comparing the system throughput under different policies, we introduce the notion of robustness of scheduling policies. Specifically, a policy that can deliver good and stable performance resilient to variations in system parameters (such as buffer sizes, processing rates, and setup times) is viewed as a “robust” policy. Numerical studies indicate that the cyclic and longest queue policies exhibit robustness in subject to parameter changes. This could provide production engineers a guideline in operation management.

Optimal production plan for a multi-products manufacturing system with production rate dependent failure rate

This study deals with the problem of dependence between production and failure rates in the context of a multi-product manufacturing system. It provides an answer about how to produce (i.e. the production rates) and what to produce (i.e. which product) over a finite horizon of H periods of equal length. We consider a single randomly failing and repairable manufacturing system producing two products Pa and Pb . The product Pa is produced to supply the strategic demand d(k) of the principal customer via a buffer stock S over k periods (k = 1, 2, … , H). The second product Pb is produced to meet a secondary but very profitable demand. It is produced during a given interval at the end of each period k. We develop a genetic algorithm to determine simultaneously the optimal production rate of the first product during each period k and the optimal duration of the production interval of the second product, maximising the total expected profit.

An Improved Genetic Approach to Optimal Supplier Selection and Order Allocation with Customer Flexibility for Multi-Product Manufacturing

As the global market becomes more competitive, manufacturing industries face relentless pressure caused by a growing tendency of greater varieties of products, shorter manufacturing cycles and more sophisticated customer requirements. Efficient and effective supplier selection and order allocation decisions are, therefore, important decisions for a manufacturer to ensure stable material flows in a highly competitive supply chain, in particular, when customers are willing to accept products with less desirable product attributes (e.g., color, delivery date) for economic reasons. This paper attempts to solve optimally the challenging problem of supplier selection and order allocation, taking into consideration the customer flexibility for a manufacturer producing multi-products to satisfy the customers’ demands in a multi period planning horizon. A new mixed integer programming model is developed to describe the behavior of the supply chain. The objective is to maximize the manufacturer’s total profit subject to various operating constraints of the supply chain. Due to the complexity and non-deterministic polynomial-time (NP)-hard nature of the problem, an improved genetic approach is proposed to solve the problem optimally. This approach differs from a canonical genetic algorithm in three aspects: a new selection method to reduce the chance of premature convergence and two problemspecific repair heuristics to guarantee feasibility of the solutions. The results of applying the proposed approach to solve a set of randomly generated test problems clearly demonstrate its excellent performance. When compared with applying the canonical genetic algorithm to locate optimal solutions, the average improvement in the solution quality amounts to as high as ten percent.