Modifying Prop Fair Scheduling Code on Vienna SL 1.9 Simulator

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #9394
    ahmetemir
    Participant

    Hello,

    I’m new on Vienna Downlik System Levem Level Simulator 1.9 Q2-2016. I want to modify prop fair scheduling algorithm to implement generalized proportinal fair scheduling algorithm. First af all I want to understand propFairSunScheduler.m of original Vienna 1.9 SL project. I read LTE-A System Level Documentation v1.9 Q2-2016 documentation. According to this documentation , proportional fair scheduler has been written , as in paper ”Reduced-Complexity Proportional Fair Scheduling for OFDMA Systems” by Z. Sun, C. Yin, and G. Yue ( link:http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4064107).

    In this paper proposed algorithm has been given on secton IV proposed PF Scheduling Algortihm For OFDMA Systems by Step1 ,Step 2 and Step 3 equations. In propFairSunScheduler.m I saw below code on PF_Scheduler function :

    metric(res(1:sum(RB_set)),:) = cleaned_c_log_matrix(res(1:sum(RB_set)),:)+UE_avgd_pre_metric_mat(res(1:sum(RB_set)),:);

    Is this code calculate PF metric? I couldn’t find any correlation between this code and step1 on this document at this link:http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4064107 . Where is the code part of Step1 ,Step2 and Step3 on original Vienna SL Simulator1.9.

    Could you give me advices?

    Best Regards…

    • This topic was modified 3 years, 2 months ago by ahmetemir.
    • This topic was modified 3 years, 2 months ago by ahmetemir.
    #9523
    ahmetemir
    Participant

    While I’m trying to solve , in below code:

    metric(res(1:sum(RB_set)),:) = cleaned_c_log_matrix(res(1:sum(RB_set)),:)+UE_avgd_pre_metric_mat(res(1:sum(RB_set)),:);`

    I realized that UE_ avd_pre_metric_ mat is calculated as:

    UE_avgd_pre_metric = -alpha_temp*log10(max(avgd_UE_throughputs+sum(RB_UEs.*c,1)*12*7,eps));
    UE_avgd_pre_metric_mat = UE_avgd_pre_metric(ones(1,N_RB),:);

    Also I’m thinking UE_avgd_pre_metric is similar as Equation(11) in the document (link: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4064107) İs it true?

    But I couldn’t understand to calculate PF metric why cleaned_c_log_matrix is added to UE_avgd_pre_metric ? What are cleaned_c_log_matrix and UE_avgd_pre_metric exactly?

    Would you explain me please? I couldn’t understand this expression from code? In order to modify prop fair scheduling I have to understand Vienna1.9 codes in detail .

    Best wishes..

    • This reply was modified 3 years, 2 months ago by ahmetemir.
    • This reply was modified 3 years, 2 months ago by ahmetemir.
    • This reply was modified 3 years, 2 months ago by ahmetemir.
    • This reply was modified 3 years, 2 months ago by ahmetemir.
    • This reply was modified 3 years, 2 months ago by ahmetemir.
    #9842

    Dear ahmetemir,

    Step 1 is represented by parameter metric, step 2 by maximizing the metric, and step 3 is represented by the loop over the number of resource blocks.

    Since the implementation was done long time ago, by another member of our team, hard to say one-to-one which parameter presents which part of the equation. To give you an idea, the parameter c represents the rate r, nominator of equation (16), while avgd_UE_throughputs represents the term (t_c – 1)*T_k.

    In order to speed up the calculation, a lot of precalculation is done and vector/matrix notation is used.

    Best regards,
    Fjolla

Viewing 3 posts - 1 through 3 (of 3 total)
  • You must be logged in to reply to this topic.