How to include some column-wise calculations in a for-loop that is otherwise working its way through a matrix row-wise?

In the code below for myFunction(), I'm trying to expand it to include calculations along the same row of the function-generated matrix where a for-loop is otherwise performing column-wise calculations. Any amounts in the "Draw_..." columns of the rendered matrix have to be reduced by the values in the columns to the left along the same row that also begin with "Draw...". Any ideas for how to do this?

Note that in the full code this example is extracted from, the number of "Draw_..." items vary based on function parameters so the references to Draw_Cl_A, Draw_Cl_B, etc., can't be hardwired but must be dynamic as the below code provides.

The outputs below shows what the code currently generates and what I would like to it to generate:

Current output:

     Inflow Target Top_up Draw_Cl_A Draw_Cl_B Draw_Cl_C
[1,]    3.6   12.0    3.6       1.0       3.0       0.5
[2,]    2.5    9.0    2.5       2.0       2.0       0.6
[3,]  160.0    6.4   15.5       2.9       2.9       2.9

Desired output:

     Inflow Target Top_up Draw_Cl_A Draw_Cl_B Draw_Cl_C  | Notes (NOT TO RENDER WHEN RUNNINGCODE!!)         |
[1,]    3.6   12.0    3.6       1.0       2.6       0.0  | Row 1: Draw_Cl_B and Draw_Cl_B change from above |
[2,]    2.5    9.0    2.5       2.0       0.5       0.0  | Row 2: Draw_Cl_B and Draw_Cl_B change from above |
[3,]  160.0    6.4   12.5       3.0       9.5       0.0  | Row 2: Draw_Cl_B and Draw_Cl_B change from above |

Explanations for above Notes for desired outputs:

Code:

serAlloc <- c(0.6, 0.5, 0.4)
inFlow <- c(6, 5, 400) 
balance <- c(1000, 900, 800)
rsv <- list(
  Target = 0.02,
  Floor = 0.005,
  Cl = c("A", "B", "C")
)
rsvTarget <- list(
  A = c(1,2,3),
  B = c(3,2,18),
  C = c(0.5,0.6,7)
)

myFunction <- function(nbr_rows) {
  mat <- matrix(0, nbr_rows, length(rsv$Cl) + 3)
  colnames(mat) <- c(
    'Inflow',
    'Target',
    'Top_up',
    paste0('Draw_Cl_', rsv$Cl)
  )
  
  for (k in 1:nbr_rows) {
    trackBal <- balance[k]
    serPct <- serAlloc[k]
    serInflow <- inFlow[k]
    mat[k, "Inflow"] <- serPct * serInflow
    
    # calculate cumulative draws for all "cl" prior to the current row by identifying "Draw" in col name
    cum_draws_lag <- if(k == 1){0} else {sum(mat[1:k-1, grepl("^Draw", colnames(mat))])}
    
    mat[k, "Target"] <- trackBal * serPct * rsv$Target
    mat[k, "Top_up"] <- min(mat[k, "Inflow"], mat[k, "Target"] + cum_draws_lag)
    
    for (cl_name in rsv$Cl) {
      target_draw <- rsvTarget[[cl_name]][[k]]
      draw <- min(target_draw,cumsum(mat[1:k,"Target"])-cum_draws_lag)
      mat[k, paste0("Draw_Cl_", cl_name)] <- draw
    }
  }
  return(mat)
}
myFunction(3)

Here's a solution. The key part to solving this was figuring out how to isolate columns to the left of the matrix cell that is currently being calculated in the for-loop (see subCols in the code below) and using the grepl() function to further subset those columns to the left of the matrix cell into those columns starting with "Draw" in their names. The values in those subset columns (and rows) are then summed. Objects subCols and sumDraws in the code below solved this issue and tell the story. Note that there were other little fixes required from the OP code, such replacing the cumsum() with sum(), changing a reference in that cumsum() (that was changed to sum()) from "Target" to "Top-up". The nice thing about for-loops like in this example is that you can patiently walk through and run every single line of code separately; I can't imagine doing this sort of thing with lapply() or map().

Code:

myFunction <- function(nbr_rows) {
  mat <- matrix(0, nbr_rows, length(rsv$Cl) + 3)
  colnames(mat) <- c(
    'Inflow',
    'Target',
    'Top_up',
    paste0('Draw_Cl_', rsv$Cl)
  )
  
  for (k in 1:nbr_rows) {
    trackBal <- balance[k]
    serPct <- serAlloc[k]
    serInflow <- inFlow[k]
    mat[k, "Inflow"] <- serPct * serInflow
    
    # calculate cumulative draws for all "cl" prior to the current row by identifying "Draw" in col name
    cum_draws_lag <- if(k == 1){0} else {sum(mat[1:k-1, grepl("^Draw", colnames(mat))])}
    
    mat[k, "Target"] <- trackBal * serPct * rsv$Target
    mat[k, "Top_up"] <- min(mat[k, "Inflow"], mat[k, "Target"] + cum_draws_lag)
    
    for (cl_name in rsv$Cl) {
      target_draw <- rsvTarget[[cl_name]][[k]]
      subCols <- which(rsv$Cl==cl_name)+3-1 # subset columns to the left of current column
      sumDraws <- sum(mat[k,1:subCols][grepl("^Draw", colnames(mat)[1:subCols])]) # sum "Draws" of cols to the left
      draw <- min(target_draw,sum(mat[1:k,"Top_up"])-cum_draws_lag - sumDraws)
      mat[k, paste0("Draw_Cl_", cl_name)] <- draw
    }
  }
  return(mat)
}
myFunction(3)