Fantahun M. Defersha

Jobshop lot streaming with routing flexibility, sequence-dependent setups, machine release dates and lag time

Lot streaming is a technique of splitting production lots into smaller sublots in a multi-stage manufacturing system so that operations of a given lot can overlap. This technique can reduce the manufacturing makespan and is an effective tool in time-based manufacturing. Research on lot streaming models and solution procedures for flexible jobshops has been limited. The flexible jobshop scheduling problem is an extension of the classical jobshop scheduling problem by allowing an operation to be assigned to one of a set of eligible machines during scheduling. In this paper we develop a lot streaming model for a flexible jobshop environment. The model considers several pragmatic issues such as sequence-dependent setup times, the attached or detached nature of the setups, the machine release date and the lag time. In order to solve the developed model efficiently, an island-model parallel genetic algorithm is proposed. Numerical examples are presented to demonstrate the features of the proposed model and compare the computational performance of the parallel genetic algorithm with the sequential algorithm. The results are very encouraging.

Simultaneous Balancing, Sequencing, and Workstation Planning for a Mixed Model Manual Assembly Line using Hybrid Genetic Algorithm

Abstract Balancing and sequencing are two important challenging problems in designing mixed-model assembly lines. A large number of studies have addressed these two problems both independently and simultaneously. However, several important aspects such as assignment of common tasks between models to different workstations, and minimizing the number and length of workstations are not addressed in an integrated manner. In this paper, we proposed a mixed integer linear programming mathematical model by considering the above aspects simultaneously for a continuously moving conveyor. The objective function of the model is to minimize the length and number of workstations, costs of workstations and task duplications. Since the proposed model cannot be efficiently solved using commercially available packages, a multi-phased linear programming embedded genetic algorithm is developed. In the proposed algorithm, binary variables are determined using genetic search whereas continuous variables corresponding to the binary variables are determined by solving linear programming sub-problem using simplex algorithm. Several numerical examples with different sizes are presented to illustrate features of the proposed model and computational efficiency of the proposed hybrid genetic algorithm. A comparative study of genetic algorithm and simulated annealing is also conducted.

Linear programming assisted (not embedded) genetic algorithm for flexible jobshop scheduling with lot streaming

Developed hybrid sequential GA which is superior to pure parallel GA.Detailed the common and distinct features of the hybrid GA and the pure GA.Distinguished two levels of hybridization of metaheuristics with Linear programming.Evaluated two alternative implementation strategies in hybridizing GA with LP. The hybridization of metaheuristics with other techniques for optimization has been one of the most interesting trends. The focus of research on metaheuristics is also becoming problem oriented rather than algorithm oriented. This has led researchers to try combining different algorithmic components in order to design more powerful algorithms. In this paper, we developed a linear programming assisted genetic algorithm for solving a flexible jobshop lot streaming problem. The genetic algorithm searches over both discrete and continuous variables in the problem solution space. A linear programming is used to assist the genetic algorithm by further refining promising solutions in a population periodically through determining the optimal values of the continuous variables corresponding to those promising solutions. This is different from one common way of hybridization referred to as linear programming embedded metaheuristics where the algorithm searches only over the integer variables and a linear programming subproblem is solved corresponding to every solution visited, which can be computationally prohibitive. Numerical examples showed that the proposed linear programming assisted (not embedded) genetic algorithm is superior to the embedded approach and as well as to a resource intensive multi-population pure parallel genetic algorithm.

A Case Study on Target Cost Estimation Using Back-Propagation and Genetic Algorithm Trained Neural Networks

Cost estimation of new products has always been difficult as only few attributes will be known. In these situations, parametric methods are commonly used using a priori determined cost function where parameters are evaluated from historical data. Neural networks, in contrast, are non-parametric, i.e., they attempt to fit curves without being provided a predetermined function. In this article, this property of neural networks is used to investigate their applicability for cost estimation of certain major aircraft subassemblies. The study is conducted in collaboration with an aerospace company located in Montreal, Canada. Two neural network models, one trained by the gradient descent algorithm and the other by genetic algorithm, are considered and compared with one another. The study, using historical data, shows an example for which the neural network model trained by genetic algorithm is robust and fits well both the training and validation data sets.

Product cost estimation using supper efficiency data envelopment analysis

The various components of parametric and non-parametric cost estimation methods assume that the historic data used in cost analysis are true representation of the relation between the cost drivers and the corresponding costs of products. However, because of efficiency variations of the manufacturers and suppliers, changes in supplier selections, market fluctuations, and several other reasons, certain costs in the historic data may be too high whereas other costs may represent better deals for their corresponding cost drivers. Thus, it may be important to rank the historic data and identify benchmarks and estimate the target costs of the product based on these benchmarks. In this paper, a novel adaptation of cost drivers and cost data is introduced in order to use data envelopment analysis for the purpose of ranking cost data and identify benchmarks, and then estimate the target costs of a new product based on these benchmarks. An illustrative case study has been presented for the cost estimation of landing gears of an aircraft manufactured by an aerospace company located in Montreal, CANADA.