AN IMPROVED BINARY INTEGER PROGRAMMING MODEL OF THE BAND COLLOCATION PROBLEM WITH LIMITED RESOURCES
Conference paper > Full Text · 2024
Authors: Ahmet Tekin, Arif GÜRSOY
VII. INTERNATIONAL HALICH CONGRESS ON MULTIDISCIPLINARY SCIENTIFIC RESEARCH, Türkiye, İstanbul, International, Verbal, Full Text, 2024-2024, pp 1220-1229
Summary
<p>When considering the telecommunication sector as a whole, the detection of potential future complexities and the reduction of potential costs are quite important. Failure to detect complexities and an increase in costs can lead to significant problems. Fiber optic is a specialized communication technology that enables data transfer by allowing light to reflect and propagate on specific glass or plastic fibers. Instead of traditional metal cables, it conducts information transmission using light signals. There are many systems that facilitate information transfer, and these systems are highly important in data transmission. Dense Wavelength Division Multiplexing (DWDM) is an advanced communication technology used in fiber optic communication. DWDM is a system that can transmit numerous data streams simultaneously over a single fiber optic cable and separates data using different wavelengths. The Band Collocation Problem (BCP) is a combinatorial optimization problem used in fiber optic communication networks in telecommunications and has a potential application area. BCP aims to pack information flows from different wavelengths in the most optimal way, aiming to achieve the minimum possible cost alongside DWDM technology. Mathematical models often assume unlimited resource usage. However, in real-life problems, having unlimited resources is not possible, which is why resource constraints become crucial. Therefore, the aim was to improve the integer linear programming model of the problem to better utilize resources. The improved model was solved using NEOS solvers. Different examples from the BCP library were used in the study, seeking solutions in different solvers (CPLEX, Gurobi, COPT). It appeared that the best solutions in terms of time and cost were obtained from the Gurobi solver. This allowed for a comparison of solvers for the problem. The results were analyzed by comparing the improved model with the old model. It was observed that the improved model provided faster and better results. The placement of information packets sent through fiber optic cables is of significant importance. Technological advancements in fiber optic networks and the decreasing functionality over time indicate the need for continued work in this field. In the future, working with larger datasets is targeted along with new approaches.</p>