Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
SES 1.2: Production Planning and Scheduling
Time:
Tuesday, 27/Jun/2017:
11:20am - 1:00pm

Session Chair: Sang Won Yoon
Location: Aula N (first floor)

Show help for 'Increase or decrease the abstract text size'
Presentations

129. Solving a multi-periods job-shop scheduling problem using a generic decision support tool

Cristovao Silva, Nathalie Klement

ENSAM, France

In this paper we present a generic decision support tool which was developed to solve many different planning problems. The proposed tool consists of a hybridization of a metaheuristic and a list algorithm. The metaheuristics can be used without any changes independently of the problem to be solved. The list algorithm must be adapted according to the studied problem. Thus, the proposed tool can support the decision process for several different planning problems with a minimum development work.

The described decision support tool was already tested with two different planning problems: (1) an activities planning and resources assignment problem in a multi-place hospital context and (2) a lot-sizing and scheduling problem with setups and due dates, for a plastic injection company. In both cases good results were obtained with the proposed tool.

In this paper we intend to present the developed tool and to describe its application to a new problem, a multi-periods job-shop scheduling, proposed by a case study company which produce industrial refrigeration equipment’s.

In the case study company, a set of metallic components are to be produced to satisfy the demand of an assembly line. Each component has to follow a processing sequence to be produced and each operation in this sequence requires a given resource (machine). The planning horizon is a week which is divided in five periods of one day. To satisfy the demand from the assembly line, a set of different lots of components is to be produced in each day of the planning horizon. Thus, we have a set of N jobs which have to be processed on a set of M machines. Each job is defined by a sequence of operations that are associated with a particular machine. Each operation has a processing time and there is a setup time between the processing of two consecutive operations which is sequence dependent. Each job has a requested period. We consider a penalty function, composed by two parts: (1) a storage cost (earliness) if the job is produced in a period prior to the requested one, (2) a tardiness cost if the job is produced in a period after the requested one. The objective is to define the operations sequence in each machine in order to minimize the total penalty.

A list algorithm is presented to be used by the tool to solve the described problem and data from the case study company were collected to generate a test instance. The decision support tool, considering the generic metaheuristic and the developed list algorithm is tested using the generated instance. The test results are presented and the ability of the proposed tool to deal with the case study company planning problem is discussed.


29. A cardinality-constrained approach for robust machine loading problems

Giovanni Lugaresi, Ettore Lanzarone, Nicla Frigerio, Andrea Matta

Politecnico di Milano, Italy

The Machine Loading Problem (MLP) refers to the allocation of operative tasks and tools to machines. Several deterministic models have been proposed in the literature for solving the MLP. However, processing times are strongly affected by uncertainty due to a variety of sources, e.g., failures, unexpected tool breaks, and unplanned maintenance interventions. As a consequence, the quality of the solution in terms of system performance is deteriorated, as the actual behavior of the system may highly differ from expectations. Thus, appropriate robust methods should be used to overcome this issue, even though the literature on robust models is scarce.

We propose a robust formulation for the MLP, with the goal of evaluating throughput bounds in the presence of a fixed number of unfortunate events over a given planning horizon. Each event is modeled as an increase of the actual processing time with respect to the nominal one, which represents the machine failure. The bounds are due to the fact that the same number of events may have a different impact on system performance, depending on how they are arranged.

The robust model proposed for the upper bound is based on the cardinality-constrained approach, in which a parameter Gamma for each tool represents the number of processing times that vary from the nominal to the maximum value due to an unfortunate event. Thus, robustness is tuned by giving a budget to the number of unfortunate events that affect each tool. The pattern of events that mostly deteriorates the system throughput is selected, and the solution is provided for this pattern, i.e., the model generates the production plan and machine tools allocations that better protect from the Gamma unfortunate events. Such robust plan can support production managers with an accurate estimation of the minimum production level that a certain system achieves in the worst conditions.

A set of realistic instances is generated to validate the robust MLP model, by tuning the size of the problem (e.g., number of produced parts, number of tools, time horizon) and the number of events in the given time horizon. The outcomes show that the objective function fairly decreases with the increase of the number of unfortunate events that affect the system. Low computational times allow the applicability in the practice. Further, this is the first application of the cardinality-constrained approach to MLP.


330. Hierarchical Sequencing of Operations with Consideration of Setups

Mayur Wakhare, Dusan Sormaz

Ohio University, United States of America

Generative Modelling methods are becoming more popular. Despite the fast and dynamic development of CAx systems, well-described procedures of Generative Model creation do not exist. The lack of the described systems and their methodologies means that only a small group of engineers have knowledge and experience to create and use such type of models. In this paper, the authors try to highlight two methods of Generative Model preparation. These methods are the results of the authors’ experiences in working with such types of models. The first method is based on cooperation with external models which are input elements into a Generative Model. Input elements (geometrical or parametrical) are one of the most important things in the process of automatic model generation. The second described method is based on an input element in a wireframe form. The paper highlights areas of application and some advantages and disadvantages for each of the presented methods.


222. Improving the Efficiency of Large Manufacturing Assembly Plants

David Sly, Michael Helwig, Guiping Hu

Proplanner, United States of America

Large manufacturing assembly plants with sub assembly lines, sequenced material deliveries, and batch driven primary manufacturing operations often struggle with coordinating their sequenced part manufacturing and kitting operations with the dynamic constraints of the main final assembly line. Additional challenges arise from the many disconnected information streams available to each group which provide delayed information with not enough part and location specific details.

Iowa State University (ISU), Proplanner and Factory Right partnered with a major Aircraft manufacturer and also a major Industrial/Ag Equipment maker to address this specific challenge with a product called Factboard. The team is being supported by the United States Army via the Digital Manufacturing and Design Innovation Institute (DMDII). DMDII is a federally-funded research and development organization of UI LABS, with a goal of increasing efficiencies of factories throughout the United States.

Making improper decisions with incomplete data reduces a factory’s throughput rate, and can result in substantial inventory increases and low overall equipment effectiveness. Pilot studies of Factboard components have demonstrated 98% reductions in line stoppages due to logistics issues, 86% reductions in on-site inventory, and 50% reductions in indirect material handling labor, all while simultaneously increasing productive throughput by nearly 10%. All of this contributes to reducing operational costs and increasing the ability of the factory and its supply chain to respond faster to changes in requirements.

A key innovation of Factboard is its ability to utilize existing transactional data within the enterprise and dynamically respond to increases, or even temporary decreases, in the quantity and quality of these real-time inputs. Because companies are often not in a position to make major upfront investments in shop floor data collection, Factboard can utilize the available information and attempt to fill in the holes to provide a real-time picture of “current events” occurring within the production systems internal to, and supplying, the final assembly line.

This is accomplished by Factboard’s ability to map engineering production life-cycle management (PLM) data sets with factory-specific build schedules and real-time transactional production and logistics data to create a series of information-rich and visually effective views designed around the needs of shop floor personas (user-defined dashboard views of production). Factboard’s decision support engine then provides specific calculations and probabilistic recommendations about inventory and resource availability at multiple points within the production system.

This paper and presentation will outline the high level data model, workflow and use-case scenarios of how the Factboard system integrates into the factory’s engineering and transactional data sources as well as how users have been able to use this more accurate, detailed and timely information to make better decisions.


189. A Heuristic Algorithm to Balance Workloads of High-Speed SMT Machines in a PCB Assembly Line

Tian He, Debiao Li, Sang Won Yoon

State University of New York at Binghamton, United States of America

-



 
Contact and Legal Notice · Contact Address:
Conference: FAIM 2017
Conference Software - ConfTool Pro 2.6.110
© 2001 - 2017 by H. Weinreich, Hamburg, Germany