Supporting data of graph representations for programmable photonic circuits

Citation Author(s):
Xiangfeng
Chen
Ugent - imec
Submitted by:
Xiangfeng Chen
Last updated:
Tue, 05/17/2022 - 22:17
DOI:
10.21227/t001-kr39
Data Format:
*.png;*.mpg
Research Article Link:
Graph Representations for Programmable Photonic Circuits
License:
Creative Commons Attribution
322 Views
Categories:
Standards Research Data
Communications
Computational Intelligence
Other
Keywords:
reconfigurable photonics; programmable photonics; large-scale photonic circuits;silicon photonics; graph theory; optical routing; optical distributions; optimal network configuration; optical filter place and routing
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

This DataPort contains supporting data of paper entitled Graph Representations for programmable photonic circuits for the Journal of Lightwave Technology. All unreachable nodes for the first graph are tested in the undirected graph with arbitrary negative weight. Unphysical paths are found on all 21 nodes. These same terminal nodes are validated in a directed graph with and without artificial nodes representing the same optical port overlapping each other. An example congestion negotiation is built on top of the single input-output pair case for 6 pair edge-disjointed shortest paths with constraints specifically for programmable photonic circuits. After several iterations with user-defined violation cost, it finds a valid solution. An animation is made based on real-time algorithm visualization on these iterations. 

Instructions: 

Folders in multipair_congestion_cost_iterations.zip:

 

1.edgeweight10541: graph with arbitrary edge weigh: bar 10, cross 5, phaseshifter 4, waveguide 1 (penalty 1 vs 2 comparison: 92 iterations vs 18 iterations, all paths total loss 550 vs 565)

 

2.edgeweight100504010: graph with arbitrary edge weigh: bar 100, cross 50, phaseshifter 40, waveguide 10 (penalty 10 vs 20 comparison: 93 iterations vs 18 iterations, all paths total loss 5500 vs 5650 ) 

 

 

Folder 1 is to support multipair inputs and outputs case in the paper. Folder 2. is further stored in Folder 4 of the next zip fie for movie demonstration. Further data for more variations are possible to demo how to design smart congestion penalties.

 

Folders in  supporting data of graph representations for programmable photonic circuits.zip:

 

1.Negative_weights_unreachable: Tested unphysical paths are found on all 21 nodes in the undirected graph with arbitrary negative weight. 

 

2.4AG_unreachable_seperated: These same terminal nodes with different light flow directions are validated in a directed graph without artificial nodes representing the same optical port overlapping each other.

 

3.4AG_unreachable_overlapped: These same terminal nodes with different light flow directions are validated in a directed graph with artificial nodes representing the same optical port overlapping each other.

 

4.movie2_6_pair_congestion_negotiation: An example congestion negotiation is built on top of the single input-output pair case for 6 pair edge-disjointed shortest paths with constraints specifically for programmable photonic circuits. After several iterations with a violation cost of 10(graph with arbitrary edge weigh: bar 100, cross 50, phaseshifter 40, waveguide 10), it finds a valid solution.  

 

5.7cells_12chose6ports_animation2: An animation is made based on real-time algorithm visualization. (red edges are congested edge pairs) 

 

 

Dataset Files

LOGIN TO ACCESS DATASET FILES

 

QUESTIONS?