Manual Task Completion Time Estimation for Job Shop Scheduling Using A Fuzzy Inference System

Manual collating and packing is still the most cost-effective way of dispatching goods in many applications, despite of the rapid development of assembly robots. One such application, is the manufacturers of Point of Sale (POS) and Point of Purchase (POP) in the design and print industry; they produce and dispatch display objects in various quantities, shapes and sizes. The display objects, typically posters and 3D displays, are designed for different commercial promotion events in supermarkets, shopping malls and other high street shops. It is difficult to assemble and pack the objects using assembly robots due to the potential complexity and infinite variety of the tasks. The collate and pack department must manually pick, collate, assemble and pack items, often carried out in multiple lines based on the nature of the jobs, as the last stage of the manufacturing process. The jobs themselves are often unique bespoke arrangements defying a generic solution, flat-packed to minimise portage costs. The design of the lines and the schedule of the lines are determined by the area manager based on their expertise and historic knowledge, which seriously limits the effectiveness of the widely available automatic global scheduling system for these POP and POS print manufacturers. This paper proposes a job completion time estimation system which estimates the completion times for different tasks under different conditions such that the intelligent scheduling system can make a schedule globally by artificially treating the assembly lines as virtual machines. The system is implemented using a particular fuzzy inference system, fuzzy interpolation, and an illustrative example demonstrates the working and potential of the proposed solution.