Tilbake til søkeresultatene

FRITEK-Teknologi

Creating robust and flexible logistics solutions using optimization

Tildelt: kr 2,6 mill.

Prosjektnummer:

171007

Søknadstype:

Prosjektperiode:

2006 - 2009

Midlene er mottatt fra:

Transportation and logistics are vital activities for the development, organization and operation of human society. Most social and economic activities involve transportation and depend upon the timely and efficient movement of people, goods and informati on. Logistics focuses on the planning and coordination of the flows of materials and information both within and between a firm and its partners. Since these activities are so important both for society and firms, quantitative planning has been standard for decades. Typical problems that require the use of quantitative tools is the setup of truck schedules for goods transporters such as Linjegods, or DHL, and the allocation of crew and physical equipment to the routes, such as determining which drivers will drive which routes and which trucks will be used where. These models are, due to both historical and practical reasons, always deterministic optimization models. The world, however, is stodchastic, and the users face a continuous stream of small and large shocks leading to interrupted operations. Hence, they also have large departments handling interruptions. However, world class scheduling and interruption departments can lead to a mediocre total behaviour, as the plans arising in deterministic mod els are neither flexible (able to adjust to shocks) nor robust (able to withstand shocks). This is not accidental, but an inherent property of deterministic models. Robustness and flexibility come at cost, and since there are no suprises in a deterministi c world, there is no reason to pay these costs. Stochastic optimization models can i principle handle flexibility and robustness, but for large cases, these models are unsolvable. We suggest a project to study, based on smaller cases, what constitutes f lexibility and robustness in some classes of discrete optimization models, and from this understanding we try to construct good solutions also for large realistic cases. This is a unique approach.

Budsjettformål:

FRITEK-Teknologi

Temaer og emner

Ingen temaer knyttet til prosjektet