Browsing by Subject "cost optimization"

Now showing 1 - 2 of 2

Access status: Open Access ,
Mechanisms for efficient inter-domain traffic distribution
(2009-07-02) (Data obrony: 2011) Kantor, Mirosław
Wydział Elektrotechniki, Automatyki, Informatyki i Elektroniki
The dissertation proposes a comprehensive approach to joint cost and performance optimization in the inter-carrier context. The research was performed for three types of proposals, which take into account cost, performance parameters and resilience aspects related to the inter-domain environment. The general mathematical formulation for the optimization problem is given together with mathematical models for four types of tariffs. Generally speaking, finding a solution for such an optimization problem could be very difficult, especially if the number of involved discrete (binary) variables is large. Thus, in order to find the solution, some heuristic algorithms are proposed. The proposed heuristics can be grouped in three sets: algorithms with a greedy-based approach, algorithms using simulated annealing and evolutionary algorithms. The set of mechanisms proposed and studied in the thesis, called Least Cost Routing (LCR) solution, form a framework which helps to optimize connections between telecommunication operators by minimizing costs for served demands and maximizing an efficient use of the existing network infrastructure. By using the results of the proposed algorithms, the routing strategy can be executed more efficiently by incorporating the knowledge of the connection cost with network conditions. Utilizing the LCR solution can also reduce time needed to analyze a huge number of alternatives and help carriers make decisions regarding new agreements with other carriers within a dynamic framework.
Access status: Open Access ,
Optimizing modular machining line design problem with mixed activation mode of machining units
(2007) Belmokhtar, Sana; Dolgui, Alexandre; Delorme, Xavier; Ignatenko, Ivan
A modular transfer line designing problem is investigated. The problem is to find the best subset of modules (machining units) from a given set and to assign them to different stations so that technological constraints and cycle upper limit are respected and the line cost is minimal. The investigated lines have a mixed activation mode for the machining units of each station, i. e. the units of each station are arranged into a series of stages such that each stage is composed of several units activated in parallel. A mixed integer program approach is proposed to model and solve the corresponding design problem. Improvements are suggested in order to reduce the model size and speed up the computations.