Jerry L. Sanders

A nesting algorithm for irregular parts and factors affecting trim losses

Abstract This paper introduces an automatic algorithm for nesting of irregularly shaped parts whose geometries were approximated by composites of non-overlapping rectangles. A performance test using randomly generated bills-of-materials (BOM) shows that the algorithm is efficient in terms of speed and material usage. This algorithm is then used to study the relationship between BOM characteristics and material trim loss. An experiment using a 25 factorial design is used to estimate the impact that the total number of parts, the number of unique part types, the part sizes (in relation to material sheet size) and the percentage of irregular parts in the BOM all will have on the material trim loss in the nesting process.

Integrating a modified simulated annealing algorithm with the simulation of a manufacturing system to optimize buffer sizes in automatic assembly systems

Selecting appropriate buffer sizes for the transport systems of automated manufacturing systems is a complex task that must account for random fluctuations in production rates by individual machines as well as the transport delays that are a part of the material handling system. If the buffer sizes are too large then the transport delays are excessive and more in-process inventories must be loaded into the system to accommodate the large buffer sizes. If the buffer sizes are too small then small processing delays will cause the buffers to fill and upstream workstations will be blocked from releasing completed workpieces. This paper integrates an extension of the simulated annealing algorithm used in combinatorial optimization with a discrete event simulation of the manufacturing system to find optimal buffer sizes for asynchronous assembly systems which involve automated inspection as well as automated assembly.

Analytical performance models for closed-loop flexible assembly systems

Flexible Assembly Systems (FASs), which form an important subset of modern manufacturing systems, are finding increasing use in todays industry. In the planning and design phase of these systems, it is useful to have tools that predict system performance for various operating conditions. In this article, we present such a performance analysis tool based on queueing approximation for a class of FASs, namely, closed-loop flexible assembly systems (CL-FASs). For CL-FASs, we describe iterative algorithms for computing steady-state performance measures, including production rate and station utilizations. These algorithms are computationally simple and have a fast convergence rate. We derive a new approximation to correct the mean delay at each queue. This improves the accuracy of performance prediction, especially in the case of small CL-FASs. Comparisons with simulation results indicate that the approximation technique is reasonably accurate for a broad range of parameter values and system sizes. This makes possible efficient (fast and computationally inexpensive) analysis of CL-FASs under various conditions.

Two-dimensional cutting stock problem research: A review and a new rectangular layout algorithm

Abstract The two-dimensional cutting stock problem addresses the allocation of a required bill of materials onto stock sheets in a manner that minimizes the trim losses. This paper surveys the progress made on the study of the problem from the original contributions by Gilmore and Gomory in the mid-1960s to the present. Conclusions are for these algorithms to find greater application in industry, one must consider the allocation of both regular and irregular shapes. Moreover, the objective of minimizing the trim losses is not an adequate performance measure when the cutting department is placed in the perspective of the entire manufacturing system. The costs of inventory and production have to be included in the objective to be minimized in order to ensure maximum production efficiency. In addition, an algorithm based on the first fit decreasing heuristic 1 is presented to achieve layouts of rectangular bills of material on rectangular stock sheets, and its performance is examined. A number of nesting algorithms are currently in use in industry, but virtually all of these systems are considered to be highly proprietary and specific. The cutting stock problem has been variously adapted to applications in numerous industries. From its original form described later, it has been modified for use in paper, lumber, cloth, metal, leather stamping and other industries. In each case, the problem has been reformulated to suit the needs of the specific industry.

Performance analysis of a LIM-based high-speed tool delivery system for machining

Material handling provides a major opportunity for productivity improvement in manufacturing systems. The poor performance of the conventional material handling systems, such as conveyors or automated guided vehicles, has led most researchers to focus on opitmal utilization of the material handling system through the design of new scheduling algorithms or control strategies. The low speed of currently available material handling systems is one of the major factors that lead to long process delays and correspondingly large work-in-process inventories common in many manufacturing systems. In this paper, we suggest a new design of material handling systems, driven by a high performance linear induction motor (LIM). The high speed of this LIM-based material handling system distinguishes it from the conventional material handling systems. To evaluate its performance, we have constructed a simulation model where the LIM-based vehicles serve as a tool delivery system in a flexible manufacturing system (FMS). We ...

Keeping pace with change: Organizational and technological imperatives

Abstract Manufacturing processes worldwide are being reconfigured by organizational innovations pioneered by the Japanese and by microelectronics-based technologies. We developed a cost model that simulated “factories” with several variations in production practices and technologies used. The results show that in a period of rapid technical change significant productivity differences can emerge between competing firms, even in “mature” sectors; these differences are amplified if learning across innovations is cumulative. The implications for developing countries are that efforts within the firm at organizational change need to be supplemented by close international relationships and appropriate infrastracture; a passive trust in the product cycle could be unhelpful.

Stochastic design optimization of asynchronous flexible assembly systems

This paper presents the application of the stochastic quasigradient method (SQG) of Ermoliev and Gaivaronski to the performance optimization of asynchronous flexible assembly systems (AFAS). These systems are subject to blocking and starvation effects that make complete analytic performance modeling difficult. A hybrid algorithm is presented in this paper which uses a queueing network model to set the number of pallets in the system and then an SQG algorithm is used to set the buffer spacings to obtain optimal system throughput. Different forms of the SQG algorithm are examined and the specification of optimal buffer sizes and pallet numbers for a variety of assembly systems with ten stations are discussed. The combined Network-SQG method appears to perform well in obtaining a near optimal solution in this discrete optimization example, even though the SQG method was primarily designed for application to differentiable performance functionals. While a number of both theoretical and practical problems remain to be resolved, a heuristic version of the SQG method appears to be a reasonable technique for analyzing optimization problems for certain complex manufacturing systems.

Chapter 5 Performance evaluation of production networks

Publisher Summary This chapter discusses different approaches to performance evaluation (PE) of discrete manufacturing systems that can be placed in three categories. These include static (allocation) models, aggregate dynamic models (ADMs), and detailed dynamic models. Except for the simplest of systems, ADMs usually provide estimates only of steady state performance measures (PMs), and often just average values are estimated (as opposed to say, entire distributions, or a 95% percentile). The PMs fall primarily into the category of physical PMs rather than financial PMs. The chapter also deals with physical PMs. Typical PMs estimated by an ADM include (all of these are steady state averages): utilization of each server; throughput (of production rate); flow time of a job through the system; queue length at a station; and work-in-process (WIP) in the system. These are among the primary physical PMs of interest to manufacturing personnel and, even though only steady state average values are estimated, they can be quite valuable in manufacturing decision making.

Modeling and design optimization of asynchronous flexible assembly systems with statistical process control and repair

This article presents the implementation of hybrid procedures involving the use of analytical performance evaluation techniques, discrete event simulation, and Monte Carlo optimization methods for the stochastic design optimization of asynchronous flexible assembly systems (AFASs) with statistical process control (SPC) and repair loops. AFASs are extremely complex and difficult to analyze in that such systems are subject to starvation and blocking effects, random jam occurrences at workstations, and splitting and merging of the assembly flow due to repair loops. Hence, an integrated approach simultaneously analyzing the interactions between product quality and optimal/near optimal system design is pursued. In the analytical analysis stage, a model based on GI/G/1 queueing network theory is used. In the Monte Carlo optimization stage, two alternative stochastic optimization approaches, namely, heuristic versions of stochastic quasigradient and simulated annealing algorithms, are implemented and compared in terms of their capabilities of solving complex AFAS design problems. The hybrid procedures presented appear to perform reasonably well in designing AFASs to reach a target production rate.

Modeling operator/workstation interference in asynchronous automatic assembly systems

The following remarks should be added after Algorithm 1 in Section 3: (1) Oj, the number of failures per unit time for machine j , is intended for situations in which Algorithm 1 will be coupled with other queueing network solution algorithms. In those situations, Oj should be recalculated during each iterative step of Algorithm 1. Steps to do this for the AAS model with interference are given in Section 4. (2) The expression derived to approximately calculate var(I) yields negative values in some special cases. A step that sets var(l) to zero in these cases should be included during implementation.