The application of AI techniques in the optimal design of multi-pass cold drawing processes

Abstract In the paper the problem of optimal pass schedule design in multi-pass wire drawing process is investigated. An automatic design procedure based on an effective artificial intelligence (AI) technique, namely simulated annealing (SA), is proposed. The developed algorithm is aimed to achieve a satisfactory balance of the drawing stresses on the material along the reduction sequence, maintaining in the meantime the drawing stress at each pass below a safety value. In this way both the optimal number of passes and the optimal drawing sequence are determined. The effectiveness of the design procedure is tested through the comparison of the sequences suggested by the algorithm with a set…

Design of the Tube Roll Forming Process Through an Heuristic Algorithm

In the paper the design of an industrial tube roll forming process is developed utilizing an heuristic algorithm, namely the simulated annealing (SA). In particular both the number and the shape of the deforming rolls are determined in order to define the forming sequence. The effectiveness of the results supplied by the SA algorithm is verified both by means of some experiments tests carried out on an industrial tube roll forming equipment and through a set of numerical simulations of the process.

Possibilistic Programming and Gas for Aggregate Production Planning under Vague Information

In production planning activities, aggregate planning is surely the phase where the lack of information is more relevant. The reasons for that can be summarised in: a) the planning product is an aggregate one, very often not corresponding to actual products; b) the delay from the actual production is relevant so that information regarding actual production resources availability and performances are vague; c) the planning activity is often based on forecasts. On the other hand, all the subsequent production planning activities are based on the result of aggregate planning. In the paper a research going on at the Workgroup for Soft Computing Application in Manufacturing is presented. A Decis…

Human factor policy testing in the sequencing of manual mixed model assembly lines

In this paper the human resource management in manual mixed model assembly U-lines is considered. The objective is to minimise the total conveyor stoppage time to achieve the full efficiency of the line. A model, that includes effects of the human resource, was developed in order to evaluate human factor policies impact on the optimal solution of this line sequencing problem. Different human resource management policies are introduced to cope with the particular layout of the proposed line. Several examples have been proposed to investigate the effects of line dimensions on the proposed management policies. The examples have been solved through a genetic algorithm. The obtained results conf…

Die design for aluminium thin tubes extrusion: a 2D simulation approach

An evolutionary approach to multi-objective scheduling of mixed model assembly lines

In this paper a multi-objective genetic algorithm for the scheduling of a mixed model assembly line is proposed, pursuing the line stop time minimisation together with the component usage smoothing. Specific features of the developed GA are step by step random selection of diversified crossover and mutation operators, population control for the substitution of duplicate chromosomes, and in-process updating of GA control parameters. Three different formulation of the fitness function were been tested with some distinct line configurations.