@rgemas Assignment Optimization Algorithm: Multi-Line Automatic Line Setup

In most production systems today the engineer balances the line in theory, while operator scheduling is either not done or not effective. When an order arrives, the engineer creates the model's operations and standard times, computes workloads against a target efficiency and hands the list to the line supervisor: "side seam needs 1.5 people". But which 1.5 people, and how efficiently they worked on that operation before, is unknown. The information stays theoretical; the supervisor builds the line from experience. When one person is absent the whole arrangement breaks, and rebuilding it is usually not done.

Arge Bilişim engineers developed a purpose-built mathematical assignment optimization algorithm that solves optimised line setup and operator scheduling simultaneously. Taking into account operation workload, model-operation similarity groups and difficulty grades, it produces the assignment that maximises line efficiency. Three modes exist: several lines for one model, several models on one line, and several models across several lines.

Without a model priority, the system treats the selected models as a single model and maximises overall efficiency. With a priority, the first model's efficiency is maximised, the preferred approach when delivery dates are close. Once the models and lines are selected and the assign button is pressed, the result reports how much of each operator's time goes to which operation and what share of each operation each operator performs, by quantity, percentage and bundle.

Real capacity. Line efficiency is calculated entirely from real data, so production planning works with dynamic, real capacity instead of a static figure. Plans fed by data replace "if we average 1,000 pieces a day the order ships Friday", and delivery delays caused by wrong capacity assumptions are prevented.

Cross-line bottleneck relief. With 15 lines and 20 models to plan, the system treats the plant as a single line and re-forms the ideal 15 lines for those 20 models. A bottleneck operation on one line is resolved with a skilled but idle operator from another.

Re-planning under absenteeism. When three operators are absent, the floor descends into chaos; the algorithm re-plans the ideal line structure for the new conditions and presents a real plan. At model changeover the supervisor is not left guessing who goes where; workload falls and the supervisor can focus on management. Line setup time, already shortened by the Bundle system, gets shorter still.

Based on past data, one operator is assigned to an operation at 41.5% efficiency, takes on 30.1% of that operation and spends 463 minutes of the day on it. The line the algorithm builds is calculated at 62.90% line efficiency. These are not estimates but real operator figures measured by argemas.