Optical Network Path Optimization With Iterative Spectrum Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to handle complex optical network structures, result in a large gap between computed service paths and optimal results, incur high costs, and require excessive computation time as network topology scales up.
Innovation Solution
A global concurrent optimization path computation method using a Path Computation Element (PCE) that iteratively optimizes service paths and spectrum resources, considering network constraints and constraints of individual services, to achieve optimal path and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If commercial solvers (CPLEX, Gurobi, XPRESS) are used to compute service paths, then linear model problems can be solved, but nonlinear model problems in complex optical network structures cannot be handled
Solution Approach 1:
The patent introduces an intermediary mechanism that transforms nonlinear optimization problems into a form that can be solved by existing linear solvers. The Path Computation Element acts as an intermediary between the complex nonlinear optical network optimization problem and the linear solvers, applying iterative algorithms to bridge the gap between problem complexity and solver capabilities.
2Measurement precision
If brute force solving method is used to compute service paths, then the best result satisfying the target function is obtained, but the computation time increases dramatically with the number of service paths
Solution Approach 1:
The patent applies partial action by implementing iterative optimization that stops when a satisfactory solution is found or when computational resources are exhausted. Rather than exhaustively searching all possible solutions (excessive action), the system performs a controlled number of iterations to achieve good enough solutions within acceptable time frames, balancing optimality with computational feasibility.
Solution Approach 2:
The patent introduces dynamic elements by making the computation process adaptive and iterative. The system dynamically adjusts its search strategy based on progress made in previous iterations, allowing it to converge toward optimal solutions more efficiently than static brute force methods while avoiding the exponential time complexity of exhaustive search.
3Adaptability or versatility
If the network topology expands and the number of service paths increases, then network coverage and capacity improve, but the time taken to compute service paths increases exponentially
Solution Approach 1:
The patent segments the large-scale optimization problem into smaller sub-problems that can be solved iteratively. By dividing the computation into manageable iterations and potentially by network segments or service groups, the system can handle expanded network topologies without requiring exponential computation time, thus maintaining productivity as network scalability increases.
4Reliability
If commercial solvers are used for path computation, then linear optimization problems can be solved, but the cost is very high charged on a per-unit-per-year basis
Solution Approach 1:
The patent replaces expensive commercial solvers with open-source alternatives (CBC, SCIP) that can be freely deployed. While these open-source solvers may require more computational effort, they eliminate the high per-unit-per-year licensing costs of commercial products like CPLEX and Gurobi, providing a cost-effective solution that maintains acceptable solution accuracy for optical network optimization.
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
A method and device for computing a global concurrent optimization path, and a computer readable storage medium. The method comprises: a path computation element (PCE) sequentially computes an actual path of each service in one or more services on the basis of a topological structure as an actual best path; the PCE sequentially allocates actual spectrum resources for the actual paths as actual best spectrum resources; the PCE computes actual target values of all services according to a target function as actual best target values; the PCE reorders service subsets in the one or more services, re-computes an actual path of each service in the service subsets and reallocates a spectrum resource, re-computes actual target values of all services, updates the actual best target values, and updates the actual best paths and actual best spectrum resources, until the number of re-computations reaches a threshold and/or the differences between the actual best target values and a minimum target value of all services are less than or equal to a predetermined value, wherein each actual best target value is the minimum value among all of the computed actual target values.