Network Design Device Optimizing Cost and Power via Solver Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network design methods struggle to efficiently design network configurations that minimize power consumption and cost, particularly in large networks, due to limitations in handling various path configurations, port use modes, and computational performance, leading to impractical design times.
Innovation Solution
A network design device that adjusts the problem scale based on arithmetic performance to solve mathematical programming problems, allowing for simultaneous design of paths, devices, and port configurations within practical time frames by limiting the number of design variables and constraints, using solvers to optimize network design operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the network design considers various path configuration patterns and port use modes to minimize power consumption and cost, then the design quality and cost-saving potential improve, but the problem scale (number of design variables and constraints) increases beyond solver performance limits
Solution Approach 1:
The patent segments the network design problem into multiple independent mathematical programming problems by dividing the set of traffic demands into subsets. Each subset is designed separately with its own constrained number of design variables and constraints, making each sub-problem solvable within practical time limits while collectively achieving comprehensive network optimization
Solution Approach 2:
The patent applies partial action by designing only a subset of traffic demands simultaneously rather than all demands at once. This partial design approach keeps the problem scale within solver performance limits while still achieving meaningful optimization for the selected subset, which can be iteratively extended to cover all demands
2Manufacturing precision
If the network design includes all traffic demands simultaneously to achieve minimum total cost, then the overall optimization improves, but the design time becomes impractical for large networks
Solution Approach 1:
The patent divides the complete set of traffic demands into multiple subsets and solves separate mathematical programming problems for each subset. This segmentation reduces the computational complexity and design time for each individual problem while collectively achieving comprehensive network optimization across all demands
Solution Approach 2:
The patent employs periodic action by iteratively solving multiple smaller design problems for different subsets of traffic demands. Each iteration processes a manageable subset, and the results are accumulated to achieve the overall optimal design, making the total design process practical for large networks
3Adaptability or versatility
If conventional labor-based network design methods are used, then flexibility in considering various design elements is maintained, but the ability to handle large numbers of parameters and dependencies is insufficient
Solution Approach 1:
The patent replaces manual labor-based network design with an automated mathematical programming system. This substitution enables the handling of large numbers of parameters and design elements through systematic computational methods, significantly increasing productivity while maintaining design flexibility through configurable objective functions and constraints
Data Source
AI summary
According to the information about a request bandwidth of a path newly requested between a starting site and an ending site of the requested path and the information about the type of a device candidate which may be provided at each site and the number of ports of each device, an objective function indicating a device cost and a constraint to be considered when the objective function is minimized are set. According to the arithmetic performance of a solver for executing the design, a problem scale corresponding to one path accommodation designing operation is set. Based on the set problem scale, a mathematical programming problem for minimizing the objective function is generated. An accommodation designing process is repeated on a plurality of paths until a solution to the mathematical programming problem is obtained by the solver, and all requested paths are completely designed.