Routing and Wavelength Assignment in WDM Optical Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current min-RWA heuristics for wavelength division multiplexing (WDM) optical networks are inefficient due to high processing times, with reported running times of up to 8 minutes for large instances, primarily due to slow shortest path queries and arc removal operations.
Innovation Solution
Implementing a method that uses breadth-first search to determine hop-count shortest paths in constant time O(m) and represents the graph using an adjacency list with doubly-linked lists for efficient arc removal, along with updating shortest path graphs and trees using Ramalingam and Reps methods to improve query and update times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If classical bin packing heuristics (FF-RWA, BF-RWA, FFD-RWA, BFD-RWA) are used for routing and wavelength assignment, then wavelength minimization is achieved, but processing time becomes excessively high (up to 8 minutes for large instances)
Solution Approach 1:
The patent segments the computation into two distinct phases: (1) routing phase that determines the path for each lightpath, and (2) wavelength assignment phase that assigns wavelengths based on the routed paths. This segmentation allows the routing to be performed once using efficient algorithms, avoiding repeated shortest path computations during wavelength assignment, thereby reducing overall processing time while maintaining wavelength minimization goals.
Solution Approach 2:
The patent performs preliminary routing of all lightpaths before wavelength assignment. By pre-computing the routes using efficient algorithms (such as constrained shortest path algorithms) and fixing these routes before the wavelength assignment phase, the system avoids re-computing shortest paths during the bin packing process. This preliminary action significantly reduces processing time while still achieving near-optimal wavelength utilization.
2Reliability
If repeated shortest path queries are performed during bin packing operations, then lightpaths can be properly routed and assigned to wavelengths, but computation time increases significantly
Solution Approach 1:
The patent extracts the shortest path computation from the iterative bin packing loop. Instead of performing shortest path queries repeatedly during each bin packing operation, the system pre-computes routes using constrained shortest path algorithms and extracts these fixed routes for use in the wavelength assignment phase. This extraction eliminates redundant computations while maintaining routing correctness.
Solution Approach 2:
The patent performs preliminary routing computations using constrained shortest path algorithms before the wavelength assignment phase. By pre-determining the routes and fixing them before bin packing operations begin, the system avoids repeated shortest path queries during wavelength assignment, thereby maintaining routing correctness while dramatically improving computation speed.
3Reliability
If arcs are removed from the graph during bin packing to prevent reuse, then wavelength assignment correctness is maintained, but graph update operations become computationally expensive
Solution Approach 1:
The patent creates a copy of the network graph for the wavelength assignment phase, using the pre-computed routes rather than dynamically modifying the original graph during bin packing. This copying approach maintains wavelength assignment correctness by preserving the route structure while avoiding expensive graph update operations. The bin packing algorithm operates on this static copied graph with fixed routes, eliminating the need for complex arc removal and graph reconstruction operations.
Data Source
AI summary
A method for routing and wavelength assignment (RWA) in an optical network with improved heuristics for reducing the computational times required for the RWA. The method minimizes the number of wavelengths by packing the lightpaths using a minimum number of bins in a bin packing problem. Computational efficiency is enhanced by using several novel methods to determine shortest paths and eliminate arcs in a graph that represents the network topology.