Minimum-cost planning of the multimodal transport of pipes with evolutionary computation
Every day many kilometres of European highways are blocked by traffic jams. Congestion on roads and at airports adds the EU's fuel bill with a corresponding rise in pollution levels. In short, our present patterns of transport growth are unsustainable. One way of easing road congestion is to...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
EDP Sciences
2009-07-01
|
| Series: | International Journal for Simulation and Multidisciplinary Design Optimization |
| Subjects: | |
| Online Access: | https://www.ijsmdo.org/articles/smdo/pdf/2009/03/smdo2009015.pdf |
| Summary: | Every day many kilometres of European highways are blocked by traffic jams. Congestion on roads and at airports adds
the EU's fuel bill with a corresponding rise in pollution levels. In short, our present patterns of transport growth are unsustainable.
One way of easing road congestion is to develop the efficient end-to-end movement of goods using two or more forms of transport in an
integrated transport chain. We will focus on the multimodal transport problem that involves finding the most economical route in the
distribution of cast iron ductile piping with or without mortar joint, both of different diameters and from different possible supply points
to different points of destination over three transport networks, road, rail and sea, which may have routes in common. The orders are made for
quantities in linear metres of pipes. The economic cost of the transport on the various routes is dependent on the number of lorries, freight
wagons and platforms required, and as these quantities must be obviously integer numbers. In practical applications the search space is
dimensionally very high, often there exist attractors into the search space due to the existence of multiple global optimum solutions,
the cost function has discontinuities, many constraints, etc. The problem that has to be resolved is of great complexity, even using
evolutionary algorithms. Our experience gained working several years in this optimization problem is described in this paper, to highlight
that there is a need using evolutionary algorithms with certain learning ingredients and using strategies that progressively impose with
more and more severity in the evolutionary optimization process the real scenario of the complex problem to get convergence to the global
optimal solutions and simultaneously to obtain low computational cost total. |
|---|---|
| ISSN: | 1779-627X 1779-6288 |