# This is the code that calculates ranking and percentage quality measure of optimization algorithms, which was described in:
# Rafał Biedrzycki: Revisiting CEC 2022 ranking: A new ranking method and influence of parameter tuning
# https://www.sciencedirect.com/science/article/pii/S2210650224001615
# author: Rafał Biedrzycki
# license: MIT
# input data format: https://github.com/P-N-Suganthan/2022-SO-BO/blob/main/CEC2022%20TR.pdf
# According to the rules of the CEC 2022 SOBC competition, all values below 1e-8 are set here to 1e-8

# As I believe that:
# “a difference is a difference only if it makes a difference.”
#  ― Darrell Huff, How to Lie with Statistics 
# the error values are rounded to 5 significant digits, and the FES (objective functions evaluations) are rounded to tenths

# exemplary results to analyze
globalAlgRanks = list( "simplifiedTuned/onlyIDEcorrItMaxWithoutSortsNoIDEbdmutNoRandxr3_3per"=0,
                       "CEC2022_EA4eig_CORRECTED/CEC2022_EA4eig_CORRECTED"=0
)

algNames=names(globalAlgRanks)
NUM_OF_ALGS=length(globalAlgRanks)

#values for CEC2022
MAX_FUN_NMBR = 12
NMBR_OF_RUNS = 30
DIMS=c(10,20)
MaxFEStab = c(2e5,1e6)


STOP_FIT = 1e-8
NEXT_LEVEL_WEIGHT = 0.01 #weight for less important criterion

SINIF_DIGITS = 5 #for rounding error. signif(12345,3)==12300; signif(0.000000012345,3)==0.0000000123
SINIF_DIGITS4FES = -1 #for rounding FES. round(12345, -1)==12340;


NUM_OF_THRES=51 #create thresholds
thresholds<- 10^seq( log10( 1e3), log10( 10^-8 ), length.out= NUM_OF_THRES)

globAlgMarks = list()
globAlgMarks2Disp = list()
algMarks4dim= list()
algMarks4dim2Disp= list()

algMarks4fun= list("10"=list("1"=0.0), "20"=list("1"=0.0))
algMarks4fun2Disp= list("10"=list("1"=0.0), "20"=list("1"=0.0))

for( ALG_NR in 1:length(globalAlgRanks)){
  algName=algNames[ALG_NR]     
  
  numOfGlobOptFound4alg = 0
  numOfThresAcchieved4alg = 0
  numOfUsedEvals4alg = 0
  perOfFreeEvals4alg = 0
  for( DIM in DIMS ){
    if(DIM==10){
      MaxFES = MaxFEStab[1]
    }else{
      MaxFES = MaxFEStab[2]
    } 
    numOfGlobOptFound4dim = 0
    numOfThresAcchieved4dim = 0
    numOfUsedEvals4dim = 0
    for( funNmbr in 1:MAX_FUN_NMBR ){
      bests4runs  =vector( mode="numeric", length=NMBR_OF_RUNS )
      stops4runs  =vector( mode="numeric", length=NMBR_OF_RUNS )
      res=read.table(paste0(names(globalAlgRanks)[ALG_NR], "_",funNmbr, "_",  DIM, ".txt"))
      res[res<STOP_FIT] = STOP_FIT
      
      stops4runs=round(res[17,], SINIF_DIGITS4FES)
      stops4runs[stops4runs>MaxFES]=MaxFES
      bests4runs = signif(res[16,], SINIF_DIGITS) 
      stops4runs[bests4runs>STOP_FIT] = MaxFES
      
      numOfGlobOptFound4fun = length(bests4runs[bests4runs==STOP_FIT])
      res4fun = numOfGlobOptFound4fun/NMBR_OF_RUNS*100
      
      numOfThresAcchieved4fun=0
      for( thresh in thresholds){
        numOfThresAcchieved4fun=numOfThresAcchieved4fun + sum(bests4runs<=thresh)
      }
      perOfAchThres4fun = numOfThresAcchieved4fun/(NMBR_OF_RUNS*length(thresholds))*100
      res4fun = res4fun+perOfAchThres4fun*NEXT_LEVEL_WEIGHT
      
      numOfUsedEvals4fun = sum(stops4runs)
      perOfFreeEvals4fun = 100-numOfUsedEvals4fun/(MaxFES*NMBR_OF_RUNS)*100
      res4fun = res4fun+perOfFreeEvals4fun*NEXT_LEVEL_WEIGHT*NEXT_LEVEL_WEIGHT
      
      numOfGlobOptFound4dim = numOfGlobOptFound4dim + numOfGlobOptFound4fun
      numOfThresAcchieved4dim = numOfThresAcchieved4dim + numOfThresAcchieved4fun
      numOfUsedEvals4dim = numOfUsedEvals4dim + numOfUsedEvals4fun

      algMarks4fun[[as.character(DIM)]][[as.character(funNmbr)]][algName] = res4fun
    	algMarks4fun2Disp[[as.character(DIM)]][[as.character(funNmbr)]][algName] = paste(round(numOfGlobOptFound4fun/NMBR_OF_RUNS*100), round(perOfAchThres4fun), round(perOfFreeEvals4fun))
    }#for( funNmbr
    res4dim = numOfGlobOptFound4dim/(NMBR_OF_RUNS*MAX_FUN_NMBR)*100
    perOfAchThres4dim = numOfThresAcchieved4dim/(NMBR_OF_RUNS*MAX_FUN_NMBR*length(thresholds))*100
    res4dim = res4dim+perOfAchThres4dim*NEXT_LEVEL_WEIGHT
    perOfFreeEvals4dim = 100-numOfUsedEvals4dim/(MaxFES*NMBR_OF_RUNS*MAX_FUN_NMBR)*100
    res4dim = res4dim+perOfFreeEvals4dim*NEXT_LEVEL_WEIGHT*NEXT_LEVEL_WEIGHT
    
    numOfGlobOptFound4alg = numOfGlobOptFound4alg + numOfGlobOptFound4dim
    numOfThresAcchieved4alg = numOfThresAcchieved4alg + numOfThresAcchieved4dim
    numOfUsedEvals4alg = numOfUsedEvals4alg + numOfUsedEvals4dim
    algMarks4dim[[as.character(DIM)]][algName] = res4dim
    algMarks4dim2Disp[[as.character(DIM)]][algName] = paste(round(numOfGlobOptFound4dim/(NMBR_OF_RUNS*MAX_FUN_NMBR)*100), round(perOfAchThres4dim), round(perOfFreeEvals4dim))
    perOfFreeEvals4alg=perOfFreeEvals4alg+perOfFreeEvals4dim
  }#for( DIM
  res4alg = numOfGlobOptFound4alg/(NMBR_OF_RUNS*MAX_FUN_NMBR*length(DIMS))*100
  perOfAchThres4alg = numOfThresAcchieved4alg/(NMBR_OF_RUNS*MAX_FUN_NMBR*length(DIMS)*length(thresholds))*100
  res4alg = res4alg+perOfAchThres4alg*NEXT_LEVEL_WEIGHT
  perOfFreeEvals4alg = perOfFreeEvals4alg/length(DIMS)
  res4alg = res4alg+perOfFreeEvals4alg*NEXT_LEVEL_WEIGHT*NEXT_LEVEL_WEIGHT
  globAlgMarks[algName] = res4alg
  globAlgMarks2Disp[algName] = paste(round(numOfGlobOptFound4alg/(NMBR_OF_RUNS*MAX_FUN_NMBR*length(DIMS))*100), round(perOfAchThres4alg), round(perOfFreeEvals4alg))
  
}#for alg
print("Results for each function")
for( DIM in DIMS ){
  for( funNmbr in 1:MAX_FUN_NMBR ){
    algMarks4funTmp = algMarks4fun[[as.character(DIM)]][[as.character(funNmbr)]]
    algMarks4fun2DispTmp = algMarks4fun2Disp[[as.character(DIM)]][[as.character(funNmbr)]]
  	ord=order(unlist(algMarks4funTmp), decreasing = T)
  	print( paste( "DIM:",DIM," fun:", funNmbr, paste(names(algMarks4fun2DispTmp[ord]), algMarks4fun2DispTmp[ord], sep = " = ", collapse = ", ")))
  }
}

print("Results for each dimensionality")
for( DIM in DIMS ){
  algMarks4dimTmp = algMarks4dim[[as.character(DIM)]]
  algMarks4dim2DispTmp = algMarks4dim2Disp[[as.character(DIM)]]
	ord=order(unlist(algMarks4dimTmp), decreasing = T)
	print( paste( "DIM:",DIM,paste(names(algMarks4dim2DispTmp[ord]), algMarks4dim2DispTmp[ord], sep = " = ", collapse = ", ")))
}

print("Final, aggregated results")
ord=order(unlist(globAlgMarks), decreasing = T)
print( paste(names(globAlgMarks2Disp[ord]), globAlgMarks2Disp[ord], sep = " = ", collapse = ", "))