Research On Sales Forecast And Production Scheduling Of Multiple Species And Small Batch

With the continuous development of the market economy, multi-variety and small batch production is taken more seriously because of its flexibility. Because of this type of production has a great variety of products and production process is very complex and the customer’s demand is likely to change, accurate production planning and scheduling is very difficult. Aiming at the problem, this thesis optimized the existing production control system start from sales forecast and production scheduling to make it more adaptable to the complex and changeable market environment.The work of this thesis is carried out from the following aspects: First, researched on sales forecasting method. The result of sales forecasting is the input of the production control system, its accuracy influences the accuracy of production planning and scheduling directly. This thesis introduced the methods and theories of sales forecasting, analyzed the advantages and disadvantages of the common forecasting methods. Because of the time-series prediction has a general applicability in commercial forecasting, the three exponential smoothing method was chose to be the sales forecasting method, to resolve the problem of a low accuracy caused by the static smoothing coefficient, designed algorithm to dynamization the smoothing coefficient to increase the accuracy of sales forecasting and made the production system could adapt the market requirement better. Second, researched on the ways of making production planning and scheduling. The traditional way of making production planning and scheduling is hierarchical, scheduling wasn’t concerned about when the production planning was making, this often causes the production planning can’t be executed and need to modify the production planning repeatedly, causes a low efficiency. In response to this issue, a method of making production planning and scheduling comprehensively was used in this thesis, a comprehensive model was proposed to describe the problem of production planning and scheduling, of which the objective contains inventory cost、shortage cost and the overtime cost, and the constraints contains the planning-requirement balance constraint、inventory constraint、processing sequence of component lots constraint、machine occupancy constraint、processing sequence between parts and components constraint、relationship of processing quantity between parts and components constraint.Finally, it is difficult to solve the discrete programming model of double layer, so the genetic algorithm was used as a tool to solve the problem. According to the characteristics of the solution contains the part of planning and the part of scheduling, the chromosome design adopts a layered coding strategy, in which the planning layer adopts real encoding, and the scheduling layer adopts symbol encoding. Considering the problem solved with multi periods and multi products, matrix encoding was adopted, and puts forward the concrete forms of chromosome design to solve the problem solution. The approach of matrix encoding was adopted considering the problem solved with multi-period and multi-product, and the concrete forms of chromosome design about this problem’s solution was put forward. At the same time, the corresponding layered decoding strategy and specific decoding methods, as well as the specific processes of hierarchical initialization, crossover, mutation and other operations were put forward. The correctness of the model and the validity of the algorithm are verified by the simulation results. The results show that, through the comprehensive formulation of production planning and scheduling, the efficiency of production scheduling was improved, and achieve the purpose of production scheduling optimization.