Industrial activities developed within competitive environments are ever more guided by a demanding and selective market, in which the efficiency in all the aspects of the production process becomes a necessary condition for the survival of the business. One of these aspects, and one which must be designed with care, is the distribution of the different activities of the production process within the plant, given its direct repercussion on production costs.; A number of procedures for the generation and evaluation of layouts have emerged since the 1950s. The fundamental differences between these are: the modelling of the problem, the number of criteria being considered, and the solution techniques employed. Within the conditions that a distribution must meet if it is to be implanted without great modification, is that the area assigned to the activities must fulfil certain conditions; it is fundamental that the size of the area be sufficient to accommodate the activity, and that its geometry permit its normal functioning. In the present work, and after an exhaustive analysis of the state of art, a method for the solution of the problem of plant layout is presented. It is a method which considers multiple criteria in the evaluation of the quality of the solutions, and which is especially efficient in the fulfilment of the geometric restrictions without losing efficiency in the consideration of the remaining criteria.; The search for the space of solutions is set out in two phases, one to limit the searching space and another of exhaustive search; each by means of genetic algorithms, which incorporate procedures to guide the exploration which have demonstrated their efficiency/effectiveness throughout the literature.; To achieve this, an analysis of the state of art was undertaken on the models and the methods of solution to the plant layout problem, both on the different approaches adopted and on the results obtained from each of the above. The state and operation of the heuristic and metaheuristic techniques employed in the problem were studied. The theory underlying evolutive algorithms was analysed and the knowledge derived thereof applied to achieve a correct and efficient application to the problem in question. An indicator on the quality of the results obtained was defined, primarily from a geometric point of view to remove the uncertainty existent in the first phase of the solution searching process. Moreover, a computer application was developed to implement the proposed algorithm, and thus enable automation of the solution to distribution problems thereby taking advantage of the calculation potential of a computer network.; In conclusion, the present work has developed an efficient method for the development of high quality plant layouts of industrial activities, whereby it is considered fundamental and inalienable the area assigned to the different activities strictly fulfils the geometric restrictions imposed.
展开▼
