User Tools

Site Tools


group:status:dsa

Table of Contents

HomeResearchMembersPublicationHonorCoursesServicesNewsOpeningsGroup

DSA Project

Personnel

  • Qiang Fan (Coordinator)
  • Jianan Bai (Co-Coordinator)
  • Lianjun Li
  • Donald
  • Emadeldin

Activities

  • 04/03/2020: Action items for next week:
    • 1. For the shadow fading, decide which option to choose, No_Yes or Yes_Yes.
    • 2. Incorporate new fading map.
    • 3. Channel allocation experiments with more channels.
    • 4. Incorporate QoS handling in the intra-channel mixing scenario.
    • 5. Update code and documentation.
  • 03/27/2020: Action items for next week:
    • 1. Check whether the fading margin is considered in both objective function and min rate constraint or not. Conduct some experiments to see the effect on different settings.
    • 2. Intra-channel mix scenario. For channel allocation, two options in step 1, a. phase II algorithm; b. only allocate 1.25MHz channels. For power allocation, incorporate intra-channel mix scenario.
  • 03/20/2020: Action items for next week:
    • 1. Phase 3 mid-April code clean up and delivery.
    • 2. Detailed documentation for phase 3 code.
  • 03/13/2020: Action items for next week:
    • 1. CPO always checks all clusters, even at the beginning not all clusters can be accommodated.
    • 2. When performing assignment check, set min rate margin and fading margin to be 0dB.
    • 3. Plot throughput vs time, use legend to differentiate priority, without and with fading margin.
    • 4. Interleave the priority.
    • 5. Phase II demo p27, rerun, 1. With different channel allocation orders (depend on priority); 2. With different fading margin settings.
    • 6. GA with sufficient condition feasibility check, option3 of slides 0228 P8.
  • 02/28/2020: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald)
  • 02/28/2020: Action items for next week:
    • 1. Incorporate two priorities/QoS/Max Power/Min rate margin into channel and power allocation.
    • 2. Incorporate the per channel max power constraint into power allocation.
    • 3. (Optional) Sufficient feasibility checking for GA channel allocation. Set max iteration for GA to save computation time.
  • 02/28/2020: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald, Emadeldin)
  • 02/21/2020: Action items for next week:
    • 1. Quantize 1d power and run the simulation for multiple rounds.
    • 2. Design GA stop criteria.
    • 3. Investigate the GA result in (2000m)^2 case, why it has lower convergence time while achieving higher throughput than other GA algorithms.
    • 4. Investigate the update ordering of DC algorithm. (see email for detail)
    • 5. Investigate channel allocation algorithms suggested by Joe. (see email for detail)
  • 02/21/2020: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald, Emadeldin)
  • 02/14/2020: Action items for next week:
    • 1. For relaxed min rate constraint in AGA, output how many cluster’s constraint has been relaxed, and what’s the final rate for those clusters.
    • 2. Simulation on GA + feasible power initialization and feasible power constraint while doing crossover and mutation, compare result with AGA.
    • 3. Provide convergence time for GA and AGA.
    • 4. Run simulations in 2000m^2 scenario, compare the performance with current 1200m^2 scenario.
    • 5. Working on simplify the min rate checking of DC algorithm.
    • 6. Working on the iterative DC algorithm.
  • 02/14/2020: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald, Emadeldin)
  • 02/07/2020: Action items for next week:
    • 1. Try bounding the power into the feasibility power region to shrink the search space.
    • 2. Try larger number of clusters in AGA for comparison. Since AGA starts with feasibility power, it should have good performance when GA fails.
    • 3. Incorporate crossover and other mutation schemes to accelerate the evolution of AGA.
  • 02/07/2020: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald, Emadeldin)
  • 01/31/2020: Action items for next week:
    • 1. Having discrete DC. After getting the optimized continuous power, round it to the nearest power level.
    • 2. Digest the slides sent by Joe. Try implementing the methods.
    • 3. For the results shown in this meeting:
      • a. Try different underlying geometry by changing the random seed. Average over multiple trials
      • b. Try including more clusters, since the runtime seems to be increasing exponentially.
  • 01/31/2020: Weekly discussion with LGS discussion materials brainstorm materials. (Lianjun, Jianan, Lingjia, Qiang Fan,Donald)
  • 01/24/2020: Phase II Demo at LGS New Jerseydemo slides. (Lingjia)
  • 01/13/2020: Weekly discussion with LGS on the phase II demo slides. Action for this week is to modify slides based on LGS suggestion by 01/17/2020.
  • 12/20/2019: Action items for next week:
    • 1. Implement DC algorithm for the new formulation. The new formulation can pave the way for multi-subcarrier setting, but current feasibility checking procedures may not work.
    • 2. Look into using genetic algorithm to solve the power allocation problem. Power can be discretized into multiple levels. Compare this with DC results.
  • 12/20/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 12/16/2019: Action items for next week:
    • 1. Try channel re-allocation based on Phase II allocation scheme.
    • 2. Implement power allocation algorithm for heterogeneous mixed channel setting.
  • 12/16/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 12/09/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 11/22/2019: Action items for next week AIs
  • 11/22/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 11/15/2019: Action items for next week AIs
  • 11/15/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 11/08/2019: Host Joe's visit to our lab.
  • 11/01/2019: Action items for next week AIs
  • 11/01/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 10/25/2019: Action items for next week AIs
  • 10/25/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 10/18/2019: Action items for next week:
    • 1. Follow option 2 in channel allocation
    • 2. Divide the code into three modules: simulation (generate the channel gains), optimization (channel allocation and power allocation based on interface information), and evaluation (throughput and coverage probability)
    • 3. Adjust system parameters (noise floor, etc.)
  • 10/18/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan)
  • 10/11/2019: Action items for next week AIs
  • 10/11/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan)
  • 10/04/2019: Action items for next week AIs
  • 10/04/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 09/27/2019: Action items for next week AIs
  • 09/27/2019: Weekly discussion with LGS discussion materials. (Lianjun, Jianan, Lingjia)
  • 09/20/2019: Action items for next week:
    • Try using minimum channel gain in minRate constraint, while still using average channel gain for objective function
    • how more details on how the power is distributed, and the cluster-wise data rate and coverage probability.
    • Modify the CA step 5 based on Joe’s suggestion: use potential system throughput instead of throughput of a single cluster.
    • Start working on baseline code for delivery.
  • 09/20/2019: Weekly discussion with LGS discussion materials. (Lianjun Jianan)
group/status/dsa.txt · Last modified: 2020/04/03 15:41 by lianjunli