Industrial Problem: Hoses Production Optimization (HoP)
The problem under investigation is about the optimization of hoses production (HoP for short) in an important Italian factory for which AntOptima made a preliminary feasibility study. The HoP is a scheduling problem that is defined by a set of real orders to be executed by a group of multi-purpose machines. Typically, for each input data there are 300 orders consisting of about 800'000 meters of hoses. Each input data specifies the production of about one month. For each order there are constraints and requirements that must be satisfied (like deadlines, raw materials demand, precedence constraints, priorities, etc.). The production must be scheduled on 35 machines with different characteristics and requirements. The goal is to organize the production without violating the "main" constraints and minimizing an objective function that takes into consideration several aspects such as the penalties of violating constraints, the customers' priorities, and so on.
In the HoP problem we have a set of customers orders. Each order is defined by the type of hose to be
produced, by the quantity, by a deadline time and by a release time (the release and the deadline define
the production time window).
Other constraints of the problems are listed below.
The goal is to choose, for each jobs Ji, a machine k of Mi and a starting time si so that the weighted combination of:
The environment implements the following search algorithms:
We implemented a Graphical environment for the management of schedules. The user can build, search and edit solutions interactively through a simple yet powerful interface.
Data is loaded from a local or remote database.
Many constraints, like machine-break-times, frozen states, time windows, ecc. can be changed at runtime.
This allows operators to simulate different scenarios and evaluate warehouses occupations.
|Network Coordinator: Marco Dorigo e-mail: firstname.lastname@example.org Web site responsibles: Christian Blum and Max Manfrin|