Carmelo J. A. Bastos-Filho

Novel routing algorithm for transparent optical networks based on noise figure and amplifier saturation

We propose and demonstrate a novel dynamic routing algorithm for transparent optical networks based on physical layer impairments such as amplifier noise accumulation, amplifier gain saturation, wavelength dependent gain and losses along lightpaths. The metrics of our algorithm is based on the calculation of the noise figure using the well known formulation for cascade of elements in lightpaths. We assume no wavelength conversion capabilities in our network and we assign wavelengths using a first fit algorithm. For a given call, our algorithm assigns the first available wavelength and calculates the best route (minimum noise figure). Upon the calculation of the noise figure of the lightpath, we obtain the bit error rate (BER). Our algorithm blocks a call if there is no wavelength available or if the BER for the available wavelength is above a given level, which guarantees a pre-defined quality-of-service for the network operation. We present simulation results of dynamic traffic in a hypothetical meshed network in terms of blocking probabilities as a function of network load and amplifier characteristics. We show that our algorithm outperforms the traditional shortest path routing algorithm.

Characterization of efficient dual-wavelength (1050 + 800 nm) pumping scheme for thulium-doped fiber amplifiers

We report on the characterization of a recently introduced dual-wavelength pumping scheme for thulium-doped fiber amplifiers using 800 and 1050 nm. Using a counterpropagating pump configuration, 180 mW of total pump power yielded 27-dB small-signal gain and less than 5-dB noise figure. Furthermore, using optical frequency domain reflectometry, the distributed gain in this configuration was evaluated, allowing for optimization of the doped fiber length.

TDFA/Raman hybrid amplifiers covering the entire S-band pumped by a single laser

Hybrid amplifier configurations are reported cascading one Raman and one or two Tm/sup 3+/-doped fiber stages, using a single pump at 1426 nm, allowing seamless amplification of the entire S-band (1460-1530 nm). Different configurations were studied, with gain over the entire band width and peak gain up to 24 dB. Noise figure as low as 7 dB was achieved.

Dual-wavelength (1050 nm +1550 nm) pumped thulium-doped fiber amplifier characterization by optical frequency-domain reflectometry

We present experimental results of distributed gain measurements from a dual-pumped (1050 nm +1550 nm) thulium-doped fiber amplifier using optical frequency-domain reflectometry. We show that significant reductions in total pump power and/or fiber length are realized with the addition of a few milliwatts at 1550 nm. For our experimental conditions, the addition of 5 mW of 1550 nm allows for a reduction of 100 mW of pump power at 1050 nm or a reduction of 44% of doped fiber length to reach the same gain as with 1050-nm pumping alone.

An evolutionary spectrum assignment algorithm for Elastic Optical Networks

A fundamental issue in Elastic Optical Networks (EONs) relies on choosing a proper route and necessary number of contiguous frequency slots from end-to-end to accommodate the traffic demands. Spectrum assignment based on the traditional First-Fit assignment has been extensively employed in EON investigations due to its inherent simplicity and favorable capacity of leaving end-to-end free slots in the network. This paper proposes a possible structure of an evolutionary algorithm that can be used to search for an appropriate spectrum ordering of the First-Fit spectrum assignment to mitigate request blocking probability. We here analyse the path request blocking probability when just lack of resource is emphasized. However, the proposal is general enough to be used in other situations.

Analysis of the Performance of the Fish School Search Algorithm Running in Graphic Processing Units

Fish School Search (FSS) is a computational intelligence technique invented by Bastos-Filho and Lima-Neto in 2007 and first presented in Bastos-Filho et al. (2008). FSS was conceived to solve search problems and it is based on the social behavior of schools of fish. In the FSS algorithm, the search space is bounded and each possible position in the search space represents a possible solution for the problem. During the algorithm execution, each fish has its positions and weights adjusted according to four FSS operators, namely, feeding, individual movement, collective-instinctivemovement and collective-volitivemovement. FSS is inherently parallel since the fitness can be evaluated for each fish individually. Hence, it is quite suitable for parallel implementations.

An adaptive-alternative routing algorithm for all-optical networks

We propose an adaptive-alternative routing algorithm for all-optical networks, named NrPSR. The algorithm finds the Nrroutes with minor cost for a given source-destination pair according to a cost function expanded in a power series (PSR) in which the coefficients are determined by a Particle Swarm Optimizer. The selected route to attend to the call request is chosen among the Nr found routes depending on a predefined policy which takes into account the wavelength availability along the links. In our simulations we considered the following physical layer effects: ASE noise generation, Optical Amplifier gain saturation, ASE saturation and OXC crosstalk. We compared the performance of our algorithm through simulation with four other well known algorithms described in the literature.

Assessment of the Power Series Routing Algorithm in Translucent, Transparent and Opaque Optical Networks

In this letter we extend the analysis of the Power Series Routing (PSR) algorithm, which was previously proposed and analyzed for transparent networks only, to the translucent and opaque networks. The PSR algorithm is a bio-inspired impairment-aware routing algorithm that can be optimized and can adapt itself to the network conditions. We demonstrate that the PSR algorithm is also suitable for these network configurations. We compared the PSR algorithm with other well known routing algorithms and we observed that the PSR algorithm achieved the best performance in all cases.

A hybrid algorithm based on fish school search and particle swarm optimization for dynamic problems

Swarm Intelligence algorithms have been extensively applied to solve optimization problems. However, some of them, such as Particle Swarm Optimization, may not present the ability to generate diversity after environmental changes. In this paper we propose a hybrid algorithm to overcome this problem by applying a very interesting feature of the Fish School Search algorithm to the Particle Swarm Optimization algorithm, the collective volitive operator. We demonstrated that our proposal presents a better performance when compared to the FSS algorithm and some PSO variations in dynamic environments.

Novel dual wavelength (1050nm + 800nm) pumping scheme for thulium doped fiber amplifiers

We demonstrate that simultaneously pumping a thulium doped fiber amplifier with 800nm and 1050nm significantly reduces the required pump powers. 180mW total pump power yielded 27dB gain and less than 5dB noise figure.