Robert T. Sumichrast

EVALUATING MIXED-MODEL ASSEMBLY LINE SEQUENCING HEURISTICS FOR JUST-IN-TIME PRODUCTION SYSTEMS

Abstract When mixed-model assembly lines use components fabricated in-house, the demand for these components is not uniform over time and is affected by the sequence of models on the assembly line. Thus, without proper mixed-model sequencing and the subsequent smoothing of component demand, the effectiveness of a just-in-time (JIT) production system is limited. This paper focuses on making component usage uniform. Five sequencing methods are reviewed, two suggested by Monden (1983) and used at Toyota (GC1 and GC2), and three proposed by Miltenburg (1989) (M-A1, MA3H1, and M-A3H2). Their performance is evaluated for the special case when all models use the same components. For all of the sequencing methods tested the mean absolute deviation of model usage varies directly with the number of models produced. There is no clear relationship between the mean absolute deviation of model usage and demand pattern or length of production sequence. Method M-A3H2 produces the highest quality feasible solutions under all conditions tested. The relative performance of the methods does not appear to be related to the number of models, demand type, or length of production sequence. To compare methods for the more general case of different models requiring different components, a mixed integer programming (MIP) model is presented as a way to find an optimal sequence. The M1P creates optimal solutions but is too slow to be used in practice. The two goal chasing heuristics used by Toyota were considered on the basis of their ability to schedule production to use components linearly over time. It is shown that these methods differ widely in their ability to generate good sequences. The performance of both is best when the products assembled have simple product structures. When models require more than one of a given component or when models require many different components, the performance worsens. The difference in performance is very small for GC1, but significant for GC2.

Evaluating sequences for paced, mixed-model assembly lines with JIT component fabrication

Evaluation measures of mixed-model assembly lines are examined. The squared deviation from linear component usage and related measures are often cited due to their apparent consistency with the goals of a just-in-time fabrication shop. This paper develops a different statistic which more directly measures deviations from linear component usage. It is shown that, for the environment studied, sequencing methods developed to achieve linear component usage, are not always better than purely random sequencing when the actual time of component use is measured. Next the sequencing procedures are compared, based on their ability to achieve desired production targets. It is shown that the sequencing method which smooths assembly times on the simulated assembly line, as well as methods which perform well based on the traditional measure of linear component use, achieve the highest levels of production. It is concluded that, in the environment studied, direct measures of system productivity are more useful than meas...

Scheduling parallel processors: An integer linear programming based heuristic for minimizing setup time

The problem of scheduling parallel processors in a make-to-stock environment with sequence setup costs is considered. A new algorithm which formulates a series of 0-1 integer sub problems is proposed and contrasted with an earlier formulation (Dearing and Henderson 1982,1984). Parallels between the sub problem formulations and generalized networks are discussed. The efficiency and quality of the solutions provided were tested using previously published data for a loom assignment problem. The heuristic solution was evaluated against the optimal integer linear programming (ILP) solution, and a rounded linear program (LP) approximation to the optimal solution for several sample problems. Results indicate that the heuristic is efficient, provides near optimal solutions to production planning problems and requires significantly less computing capability than previously reported LP, TLP approaches.

A heuristic and lower bound for a multi-depot routing problem

Abstract The multi-depot routing problem considered here deals with a situation where a fleet of τ trucks is used to transport π different raw materials available at σ sources to π plants. A heuristic is presented for solving this problem. An initial feasible solution is found by determining the least costly way of supplying each plant with the material demanded, one plant at a time. Then routes are exchanged for each truck if a net cost savings can be realized, while maintaining feasibility. The heuristic is compared to a lower bound obtained from a relaxed binary formulation. Both methods are applied for four sets of randomly generated problems with different values for τ, σ, π , and for the maximum number of trips allowed. The problems ranged in size from 52 to 609 nodes in the lower bound formulation. The results from the heuristic exceed the bound by 2–4% for problem sets with 52–303 nodes. For the larger problems this percentage reaches 8%. The gap is due partly to the lower bound having a smaller value than the optimal solution and partly to the heuristic being non-optimal.

CPG EA : a grouping genetic algorithm for material cutting plan generation

Construction firms specializing in large commercial buildings often purchase large steel plates, cut them into pieces and then weld the pieces into H-beams and other construction components. We formalize the material ordering and cutting problem faced by this industry and propose a grouping genetic algorithm, called CPGEA, for efficiently controlling the relevant costs. We test the quality of CPGEA in various ways. Three sets of simulated problems with known optimal solutions are solved using CPGEA, and the gap between its solutions and optimal solutions is measured. The same problem sets are also solved with an expert system and a multi-start greedy heuristic. CPGEA solutions are found to be consistently lower cost than the competing methods. The difference in solution quality is most pronounced for difficult problems requiring multiple identical plates in the optimal solution. CPGEA is also tested using data from actual construction projects of a company faced with this problem. Since an optimal solution for the problems is not available, a lower bound is created. For the historical problems tested, the average percent difference between CPGEA solutions and the lower bound is 0.67%. To put this performance in context, the results of solving these problems with an expert system and using experienced engineers is also reported. Of these three methods, CPGEA achieves the best performance and the human experts the worst performance.

An Evolutionary Algorithm for Sequencing Production on a Paced Assembly Line

A new sequencing method for mixed-model assembly lines is developed and tested. This method, called the Evolutionary Production Sequencer (EPS) is designed to maximize production on an assembly line. The performance of EPS is evaluated using three measures: minimum cycle time necessary to achieve 100% completion without rework, percent of items completed without rework for a given cycle time, and sequence “smoothness.” The first two of these measures are based on a simulated production system. Characteristics of the system, such as assembly line station length, assembly time and cycle time, are varied to better gauge the performance of EPS. More fundamental variation is studied by modeling two production systems. In one set of tests, the system consists of an assembly line in isolation (i.e., a single-level system). In another set of tests, the production system consists of the assembly line and the fabrication system supplying components to the line (i.e., a two-level system). Sequence smoothness is measured by the mean absolute deviation (MAD) between actual component usage and the ideal usage at each point in the production sequence. The performance of EPS is compared to those of well-known assembly line sequencing techniques developed by Miltenburg (1989), Okamura and Yamashina (1979), and Yano and Rachamadugu (1991). EPS performed very well under all test conditions when the criterion of success was either minimum cycle time necessary to achieve 100% production without rework or percent of items completed without rework for a given cycle time. When MAD was the criterion of success, EPS was found inferior to the Miltenburg heuristic but better than the other two production-oriented techniques.

Expert-System Technology for Hotels: Concierge Application

Concierge service at hotels can be augmented by installing an expert system, which makes use of artificial intelligence. Such a system can be a valuable supplement to human-concierge services and a...

A new cutting-stock heuristic for scheduling production

Abstract This study interprets a scheduling problem in the woven fiberglass industry as an example of the cutting-stock problem; where wasted production capacity rather than wasted material is to be controlled. The solution is complicated due to the need to consider setup costs, so a heuristic is developed and tested. Comparison to one companys a historical production decisions indicates that both wasted capacity and setup Costs can be substantially reduced through application of the heuristic.

Scheduling parallel processors to minimize setup time

Abstract The problem of scheduling parallel processors in a make-to-stock environment with sequencedependent setup costs is considered. A new algorithm which formulates a series of 0 1 integer subproblems is proposed and contrasted with an earlier formulation. Parallels between the subproblem formulations and generalized networks are discussed.

A multi-expert system for material cutting plan generation

Abstract Construction firms specializing in large commercial buildings must often design and build steel structural elements as a part of each project. Such firms must purchase large steel plates, cut them into pieces and then weld the pieces into H-beams and other construction components. We formalize the material ordering and cutting problem faced by this industry and propose an expert system for efficiently combining steel elements into plates to control relevant costs. This expert system is based on the procedural knowledge of multiple experts rather than on the rules of a single expert as is more common. We tested the expert system using data supplied by Lien-Kang Heavy Industrial Company, Ltd. (LK) as well as using simulated data with known optimal solutions. Comparison to LKs solutions indicates that the expert system solution is less costly in every historical project. Testing on 40 simulated projects reveals how problem parameters affect performance. The expert system has been embedded into a decision support system (DSS) that LK is currently using. The qualitative benefits of using this computerized system include a significant reduction in the time and effort required for generating a solution and increased report accuracy.