Abstract: Production is becoming more and more complex with the manufacturing system changing from mass manufacturing to multi-varieties and small-batch production. A better way to improve production efficiency is considering processing and assembly in multi-step and multi-cells within the production process. However, the present researches in discrete manufacturing systems focused on workshop or multi manufacturing cells scheduling, but rarely on the parallel collaborative scheduling of multi manufacturing cells. Moreover, the transport time, changeover time, and debugging time were ignored, which resulted in deviations between scheduling results and actually machining. In fact, changeover time, and debugging time, transport time between cells during product switching have taken place in the discrete numerical controlled manufacturing cell all the time. And the production time of spare parts fabricated in the product could not satisfy the assembly requirement. As a result, parallel collaborative scheduling when processing and assembling multi manufacturing cells workpieces for multi products is a proper way to improve the production efficiency.
To promote the performance of production, multi-manufacturing cells are regarded as a manufacturing system (MS) proposed a collaborative scheduling method based on parallel manufacturing, then a parallel collaborative scheduling model is presented by considering the transport time, changeover time, and debugging time, In order to solve the efficiently of the model, parallel collaborative genetic algorithm (GA) is also proposed. Based on this, the method is tested in a complex electromechanical products workshop which has multi- manufacturing cells, and then a comparative variable batch scheduling model and equal batch was conducted. Through the studies,the parallel collaborative scheduling can significantly reduce the production cycle and promote the performance of machine.
The research shows that parallel collaborative scheduling can make the tasks reasonably assigned to machines, shorten the assembly waiting time and the time of the tooling changing, installment and testing within aitifact swithing, so as to increase the availability of machine and shorten the production cycle. It can be seen that the parallel collaborative scheduling can significantly reduce the production cycle and promote the machine performance.

Key words: scheduling, parallel collaborative genetic algorithm, manufacturing cell