Browsing by Subject "Integer programming."

Now showing 1 - 5 of 5

Open Access
Computational analysis of the search based cuts on the multidimensional 0-1 knapsack problem
(2003) Pekbey, Duygu
In this thesis, the potential use of a recently proposed cut (the search based cut) for 0-1 programming problems by Oguz (2002) is analyzed. For this purpose, the search based cuts and a new algorithm based on the search based cuts are applied to multidimensional 0-1 knapsack problems from the literature as well as randomly generated multidimensional 0-1 knapsack problems. The results are compared with the implementation of CPLEX v8.1 in MIP mode and the results reported.
Open Access
Improving application behavior on heterogeneous manycore systems through kernel mapping
(2013) Albayrak, Ömer Erdil
Many-core accelerators are being more frequently deployed to improve the system processing capabilities. In such systems, application mapping must be enhanced to maximize utilization of the underlying architecture. Especially, in graphics processing units (GPUs), mapping kernels that are part of multi-kernel applications has a great impact on overall performance, since kernels may exhibit different characteristics on different CPUs and GPUs. While some kernels run faster on GPUs, others may perform better in CPUs. Thus, heterogeneous execution may yield better performance than executing the application only on a CPU or only on a GPU. In this thesis, we investigate on two approaches: a novel profiling-based adaptive kernel mapping algorithm to assign each kernel of an application to the proper device, and a Mixed Integer Programming (MIP) implementation to determine optimal mapping. We utilize profiling information for kernels on different devices and generate a map that identifies which kernel should run where in order to improve the overall performance or energy consumption of an application. Initial experiments show that our approach can efficiently map kernels on CPUs and GPUs, and outperforms CPU-only and GPU-only approaches. Some part of this work is published in 41st International Conference on Parallel Processing Workshops (ICPPW), 2012 [1], and submitted to Parallel Computing journal (ParCo) [2].
Open Access
A multiperiod capital budgeting problem: the ten-year acquisition program of the Turkish Armed Forces
(2000) Yavuz, Serdar
Structuring the Ten-year Acquisition Program of the Turkish Armed Forces is a critical problem faced during the defense planning process. This is essentially a resource allocation process. Since the decisions have long-term impacts on the defense capability of the Turkish Armed Forces, it is a high-priority concern of the top decision makers. In this study, a binary integer programming model is proposed which will aid the decision makers in their formulation of the acquisition process. This model determines which projects to select as well as the years at which they should be started. Viability of the model is shown by means of representative computational runs.
Open Access
Outer approximation algorithms for the congested p-median problem
(2011) Şelfun, Selva
In this thesis, we study a generalization of the p-median problem, which is a well-known facility location problem. Given a set of clients, a set of potential facilities, and a positive integer p, the p-median problem is concerned with choosing p facilities and assigning each client to an open facility in such a way that the sum of the travel times between each client and the facility that serves that client is as small as possible. The classical p-median problem takes into account only the travel times between clients and facilities. However, in many applications, the disutility of a client is also closely related to the waiting time at a facility, which is typically an increasing function of the demand allocated to that facility. In an attempt to address this issue, for a given potential facility, we define the disutility of a client as a function of the travel time and the total demand served by that facility. The latter part reflects the level of unwillingness of a client to be served by a facility as a function of the level of utilization of that facility. By modeling this relation using an increasing convex function, we develop convex mixed integer nonlinear programming models. By exploiting the fact that nonlinearity only appears in the objective function, we propose different variants of the well-known outer approximation algorithm. Our extensive computational results reveal that our algorithms are competitive in comparison with the off-the-shelf solvers.
Open Access
Some heuristics as preprocessing for 0-1 integer programming
(1991) Yılmaz, Fatih
It is well-known that 0-1 integer programming is one of the hard problems to solve other than special cases of constraint set in mathematical programming. In this thesis, some preprocessing will be done to get useful informations, such as feasible solutions, bounds for the number of Ts in feasible solutions, about the problem. A new algorithm to solve general (nonlinear) 0-1 programming with linear objective function will be devoloped. Preprocessing informations, then, are appended to original problem to show improvements in enumerative algorithms, e.g. in Branch and Bound procedures.