Escalonamento Da Produção Para Sistema De Manufatura Job Shop Com Parâmetros Inteligentes

Abstract

The development of an effective planning process for the sequence of processing orders in manufacturing systems programming is a task with a high degree of complexity. The absence of platforms for experimentation simulation of the scale of production, whatever the typology: flow shop, job shop, open shop, hampers the methodological learning curve for this type of problem. This one work proposes the design of a web application, which implements a genetic algorithm (GA) to minimize the makespan (completion time), to allow simulation of benchmark sets of production scheduling in job shop manufacturing systems. The web application is available at <http://iproductionscheduling.com> and was developed using Python and HTML5 languages. In this way, it is possible to carry out optimized simulations of the instances of the type abz, dum, ft, yn, la, orb, swv and ta; following the premise that the job emerges according to a production order issued with manufacturing schedule and time specifications with particularities contained in a benchmark set. The genetic operators (roulette crossover and mutations) were adapted to promote intensification and exploration in the search space. Elitism and random immigrants were used as a technique for controlling population diversity. In the testing phase, were tested in isolation with the abz5 10 × 10 of different variations in GA parameters in the expected result. After this, the application was evaluated from two instances, abz5 10 × 10 and ft06 10 × 10, with results compatible with those of the recent literature, obtained by other heuristic methods. At Experiments carried out proved that the algorithm implemented in the core of the page the current optimal limits and adds when it provides experimentation and shows the results of the Gantt chart, in addition to shown tables and graphs to evaluate the optimization process with the parameters determined by the user.