John O. McClain

The role of work-in-process inventory in serial production lines

In serial production systems, storage may be provided between processes to avoid interference due to lack of synchronization. This paper investigates the behavior of lines buffered in this way and explores the distribution and quantity of work-in-process WIP inventory that accumulates. We study simple, generic production systems to gain insight into the behavior of more complex systems. The authors are surprised by the sometimes counterintuitive results, but are joined in this surprise by both academics and practitioners with whom the study has been discussed. Results are presented for: identical workstations with and without buffers; balanced lines in which variability of processing times differs between stations; unbalanced lines; and lines with unreliable workstations. In general, buffers between workstations increase system capacity but with sharply diminishing returns. Position as well as capacity of the buffers are important. These results are preliminary, to be confirmed and extended by further study-indeed, a primary purpose of this paper is to stimulate such study. However, even these preliminary results yield design guidelines that should be useful in industrial practice.

Commissioned Paper: On the Interface Between Operations and Human Resources Management

Operations management (OM) and human resources management (HRM) historically have been very separate fields. In practice, operations managers and human resource managers interact primarily on administrative issues regarding payroll and other matters. In academia, the two subjects are studied by separate communities of scholars publishing in disjoint sets of journals, drawing on mostly separate disciplinary foundations. Yet, operations and human resources are intimately related at a fundamental level. Operations are the context that often explains or moderates the effects of human resource activities such as pay, training, communications, and staffing. Human responses to OM systems often explain variations or anomalies that would otherwise be treated as randomness or error variance in traditional operations research models. In this paper, we probe the interface between operations and human resources by examining how human considerations affect classical OM results and how operational considerations affect classical HRM results. We then propose a unifying framework for identifying new research opportunities at the intersection of the two fields.

Multilevel capacitated lotsizing complexity and LP-based heuristics

Abstract This paper presents the first heuristics capable of solving multilevel lotsizing problems with capacity constraints on more than one level. Moreover, the form of the heuristics is quite general so that they can easily be extended to solve a variety of problems. If one wants to solve these problems on a routine basis in a real environment one needs to find fast and easy algorithms. However, we show that for certain problem classes this is a very difficult task, far more difficult than has been suggested in the literature. For problems with setup times we show that finding a feasible solution is NP-complete. Even without setup times testing for feasibility can be very difficult. Just how time consuming such heuristics must be is demonstrated. This leaves little chance to build fast and easy heuristics except for the most simple cases. Our exploration of the complexity issues points to mathematical programming as a potential source of heuristics for these problems. This paper presents a new and general approach based on rounding an LP solution for the problem without setup times. The methods use different information and patterns evident in the LP solution are explored. The approach is tested on a large set of problems. The main contributions of this paper are the way in which we distinguish between the easy and hard lotsizing problems, the LP-based heuristics and the test set of capacitated multilevel lotsizing problems.

Overcoming the dark side of worker flexibility

Abstract Flexible work assignment has great potential to increase productivity. When bottlenecks develop, for example, downstream operations may halt for lack of materials. A flexible worker can prevent this by moving in and increasing capacity temporarily, thereby avoiding work stoppages. In practice, it is not that simple. This paper identifies several “negative side effects” that occur in systems that rely on worker flexibility, effects that may partially or totally offset the advantages. Research has shown that performance feedback and work interruptions are factors that may explain some of these effects. Behavioral theories about those two factors lead to a series of hypotheses and two experiments. The results show that productivity loss due to these behavioral effects can be significant, in both the statistical and managerial sense of the word. Implications for the design and operation of flexible work environments are discussed, including methods for mitigating or eliminating the negative side effects, resulting in a meaningful productivity gain.

On-the-fly line balancing with very little WIP

Abstract Dynamic line balancing (DLB) allows the workers on a production line to help each other with some tasks. However, for buffer capacity of 2 or less per stage, the advantage of DLB appeared to be lost. In this paper we show that DLB can increase efficiency even when buffers are absent. The improvement is a result of using more information when deciding who should do the shared task. The new information is supplied by a more careful measurement of how much unfinished work is represented by the jobs in the system. Previous results are extended to situations where jobs must be done in the order of arrival, which precludes using rules like SPT (shortest processing time) to decide which job to do next. It is shown that DLB offers improved efficiency in this case as well. Furthermore, this additional restriction makes it even more valuable to use additional information in deciding when workers should help each other.

The good, the bad and the ugly: Quality on an assembly line

SUMMARY Firms can handle defects that occur on an assembly line in a number of different ways. Traditionally, defective units were either scrapped or sent to a separate repair station. With the advent of Japanese production methodology in the US, line-stop buttons were introduced to assembly lines; when a defect was discovered, the assembly line was stopped and the defect was repaired immediately (on-line repair). We develop a relatively simple model of the assembly line to compare these alternate strategies and derive some qualitative insights into this problem. Our primary performance measures are the production throughput rate, the work-in-process inventory level (WIP), and the field defect rate. We use both queuing theory and simulation to estimate these measures for the different strategies Of course, not all defects can be detected. The presence of imperfect inspection complicates the analysis, requiring us to determine values for probabilities of defects and inspection errors that are consistent wi...

Response-Variance Tradeoffs in Adaptive Forecasting

This paper centers on what happens to an exponential smoothing forecasting system when the smoothing parameters are changed. It gives sufficient conditions under which several proposed models give numerically identical results, and notes a surprising symmetry in one basic model between the smoothing constants for the smoothed average and the smoothed trend. The paper defines smoothing-constant values that cause oscillatory behavior, discusses the difficulties caused by oscillation, and finds that commonly used values of the smoothing constants lie in a region of oscillation of varying period. It uses these results in an analytical study of the tradeoff between responsiveness and variance; some dominant choices emerge. Finally, on the basis of these results, the paper examines several sample sets of data from the literature, and discusses implications for the choice of smoothing constants in complex situations.

Dominant tracking signals

Abstract The objective of this research is to find the best methods of automatically monitoring an exponentially smoothed forecast. Three conclusions are drawn. First, previously used performance measures are inadequate. As a consequence, currently available control limits can give false alarm rates that are substantially different than advertised. Second, two commonly used tracking signals can be substantially improved by choice of smoothing parameters. Finally, when measured by the new criteria proposed in this paper, the smoothed error tracking signal is substantially better than the unweighted sum of errors (CUSUM) method.

Efficient Solutions to a Linear Programming Model for Production Scheduling With Capacity Constraints and No Initial Stock

Abstract In this paper we present a decomposition approach to solve large scale linear programming models for production scheduling when there are multiple capacity-constrained facilities. The formulation assumes that there are no initial inventories, and hence is most useful in a planning environment where the current shop status is not the primary concern. The approach can be implemented as an exact procedure or with heuristic stopping rules. We determine problem characteristics for which the decomposition approach is faster than LP, so that very large problems could be solved. Problem difficulty is found to be related to size and tightness of the capacity constraints. Quality-of-solution versus CPU time tradeoffs are given for various stopping rules. Finally, we discuss the potential importance of this formulation and approach in manufacturing problems.

Cyclic Assembly Schedules

Abstract Cyclic assembly schedules are schedules which follow a repeating production pattern. Each component is produced at times that are integer multiples of the base planning period of the finished product. These schedules have the managerial advantages of ease of control and planning, since future production follows a familiar, repeating pattern. Establishing and maintaining such a schedule requires a system to protect the schedule when bottlenecks occur. This paper discusses some of these issues and describes a simple method for calculating the optimal parameters of a cyclic schedule in a multistage assembly operation when production intervals are power-of-two multiples of a base period specified by management. The cycles are based on an Economic Order Quantity analysis, which is applicable if demand for the end items is fairly steady as in the case of a facility that supplies components to an assembly line.