vignettes/ffv_opt_solin_relow_allfn.Rmd
ffv_opt_solin_relow_allfn.RmdTest the lower-bounded linear-continuous optimal allocation solution function from package. Use the Guatemala-Cebu scrambled nutrition and height early childhood data, same as line by line.
The objective of this file is to solve the linear \(N_i\) and \(H_i\) problem without invoking any functions. This function first tested out the linear solution algorithm.
File name ffv_opt_solin_relow_allfn:
See associated functions in the ffv_opt_dtgch_cbem4 file.
Generate four categories by initial height and mother’s education levels combinations.
# Load Data and Estimation Results: A and alpha, lin and loglin
ls_opti_alpha_A <- ffy_opt_dtgch_cbem4()## Warning: The `x` argument of `as_tibble.matrix()` must have unique column names if `.name_repair` is omitted as of tibble 2.0.0.
## Using compatibility `.name_repair`.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was generated.
df_raw <- ls_opti_alpha_A$df_raw
df_hw_cebu_m24 <- df_raw
df_esti <- ls_opti_alpha_A$df_esti
# Review dataframes
# raw file
head(df_raw, 10)## # A tibble: 10 x 17
## S.country vil.id indi.id sex svymthRound momEdu wealthIdx hgt wgt
## <chr> <dbl> <dbl> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Cebu 1 2 Female 24 7.1 7.3 79.2 10776.
## 2 Cebu 1 3 Female 24 9.4 10.3 76.7 10296.
## 3 Cebu 1 4 Female 24 13.9 13.3 78.1 7604.
## 4 Cebu 1 5 Male 24 11.3 9.3 84.1 11787.
## 5 Cebu 1 6 Female 24 7.3 7.3 76.9 7991.
## 6 Cebu 1 7 Male 24 10.4 8.3 79.6 12583.
## 7 Cebu 1 8 Female 24 13.5 9.3 81.5 8358.
## 8 Cebu 1 9 Female 24 10.4 17.3 74.7 8195.
## 9 Cebu 1 10 Male 24 10.5 6.3 77.1 9442.
## 10 Cebu 1 11 Male 24 1.9 6.3 72.1 6627.
## # ... with 8 more variables: hgt0 <dbl>, wgt0 <dbl>, prot <dbl>, cal <dbl>,
## # p.A.prot <dbl>, p.A.nProt <dbl>, momEduRound <fct>, hgt0med <fct>
head(df_esti, 10)## # A tibble: 10 x 9
## S.country vil.id indi.id svymthRound alpha_lin alpha_log A_lin A_log beta
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Cebu 1 2 24 0.0101 0.00669 78.4 4.34 0.000959
## 2 Cebu 1 3 24 0.0323 0.00925 79.9 4.36 0.000959
## 3 Cebu 1 4 24 0.0323 0.00925 78.9 4.35 0.000959
## 4 Cebu 1 5 24 0.0662 0.0139 80.6 4.37 0.000959
## 5 Cebu 1 6 24 0.0101 0.00669 77.8 4.34 0.000959
## 6 Cebu 1 7 24 0.0662 0.0139 79.6 4.36 0.000959
## 7 Cebu 1 8 24 0.0101 0.00669 77.8 4.34 0.000959
## 8 Cebu 1 9 24 0.0101 0.00669 78.0 4.34 0.000959
## 9 Cebu 1 10 24 0.0662 0.0139 80.0 4.36 0.000959
## 10 Cebu 1 11 24 0.0602 0.0121 77.7 4.33 0.000959
# Attach
attach(df_raw)## The following objects are masked from df_hw:
##
## cal, hgt, hgt0, hgt0med, indi.id, momEdu, momEduRound, p.A.nProt,
## p.A.prot, prot, S.country, sex, svymthRound, vil.id, wealthIdx,
## wgt, wgt0
# Child Count
df_hw_cebu_m24_full <- df_hw_cebu_m24
it_obs = dim(df_hw_cebu_m24)[1]
# Total Resource Count
ar_prot_data = df_hw_cebu_m24$prot
fl_N_agg = sum(ar_prot_data)
# Vector of Planner Preference
ar_rho = c(seq(-200, -100, length.out=5), seq(-100, -25, length.out=5), seq(-25, -5, length.out=5),
seq(-5, -1, length.out=5), seq(-1, -0.01, length.out=5), seq(0.01, 0.25, length.out=5), seq(0.25, 0.99, length.out=5))
ar_rho = c(-50, -25, -10)
ar_rho = unique(ar_rho)This also works with any CRS CES.
# Matrixes to Store Solution Results
tb_opti_alloc_all_rho <- df_esti
mt_hev_lin = matrix(, nrow = length(ar_rho), ncol = 2)
mt_opti_N = matrix(, nrow = it_obs, ncol = length(ar_rho))
mt_opti_H = matrix(, nrow = it_obs, ncol = length(ar_rho))
# A. First Loop over Planner Preference
# Generate Rank Order
for (it_rho_ctr in seq(1,length(ar_rho))) {
rho = ar_rho[it_rho_ctr]
# B. Parameters for solving the optimal allocation problem
df <- df_esti
svr_A_i <- 'A_lin'
svr_alpha_i <- 'alpha_lin'
svr_beta_i <- 'beta'
fl_N_agg <- fl_N_agg
fl_rho <- rho
# C. Invoke optimal linear (crs) solution problem
# ar_opti is the array of optimal choices, it is in df_opti as well.
# use df_opti for merging, because that contains the individual keys.
# actually file here should contain unique keys, unique key ID as required input. should return?
# actually it is fine, the function here needs the key, not solin_flinr
svr_inpalc <- 'opti_allocate'
svr_expout <- 'opti_exp_outcome'
ls_lin_solu <- ffp_opt_solin_relow(df, svr_A_i, svr_alpha_i, svr_beta_i, fl_N_agg, fl_rho,
svr_inpalc, svr_expout)
tb_opti <- ls_lin_solu$df_opti
ar_opti_inpalc <- ls_lin_solu$ar_opti_inpalc
ar_opti_expout <- ls_lin_solu$ar_opti_expout
mt_opti_N[, it_rho_ctr] = ar_opti_inpalc
mt_opti_H[, it_rho_ctr] = ar_opti_expout
# m. Keep for df collection individual key + optimal allocation
# _on stands for optimal nutritional choices
# _eh stands for expected height
tb_opti_main_results <- tb_opti %>%
select(-one_of(c('lowest_rank_alpha', 'lowest_rank_beta')))
tb_opti_allocate_wth_key <- tb_opti %>% select(one_of('indi.id', svr_inpalc, svr_expout)) %>%
rename(!!paste0('rho_c', it_rho_ctr, '_on') := !!sym(svr_inpalc),
!!paste0('rho_c', it_rho_ctr, '_eh') := !!sym(svr_expout))
# n. merge optimal allocaiton results from different planner preference
tb_opti_alloc_all_rho <- tb_opti_alloc_all_rho %>% left_join(tb_opti_allocate_wth_key, by='indi.id')
# print subset
if (it_rho_ctr == 1 | it_rho_ctr == length(ar_rho) | it_rho_ctr == round(length(ar_rho)/2)) {
print('')
print('xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx')
print(paste0('xxx rho:', rho))
print('xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx')
print('xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx')
print(summary(tb_opti_main_results))
}
}## Warning: Unknown columns: `lowest_rank_alpha`, `lowest_rank_beta`## [1] ""
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxx rho:-50"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## S.country vil.id indi.id svymthRound
## Length:1043 Min. : 1.00 Min. : 2.0 Min. :24
## Class :character 1st Qu.: 7.00 1st Qu.: 298.5 1st Qu.:24
## Mode :character Median :12.00 Median : 627.0 Median :24
## Mean :12.52 Mean : 634.8 Mean :24
## 3rd Qu.:17.00 3rd Qu.: 939.5 3rd Qu.:24
## Max. :33.00 Max. :1349.0 Max. :24
## alpha alpha_log A A_log
## Min. :0.01014 Min. :0.006687 Min. :73.70 Min. :4.287
## 1st Qu.:0.01014 1st Qu.:0.006687 1st Qu.:77.73 1st Qu.:4.334
## Median :0.06018 Median :0.012090 Median :78.45 Median :4.343
## Mean :0.04108 Mean :0.010321 Mean :78.46 Mean :4.343
## 3rd Qu.:0.06018 3rd Qu.:0.012090 3rd Qu.:79.15 3rd Qu.:4.351
## Max. :0.06619 Max. :0.013940 Max. :84.61 Max. :4.417
## beta rank_val rank_idx lowest_rank_A
## Min. :0.0009588 Min. : 89.24 Min. : 1.0 Min. :73.7
## 1st Qu.:0.0009588 1st Qu.: 95.00 1st Qu.: 261.5 1st Qu.:73.7
## Median :0.0009588 Median : 96.39 Median : 522.0 Median :73.7
## Mean :0.0009588 Mean : 96.25 Mean : 522.0 Mean :73.7
## 3rd Qu.:0.0009588 3rd Qu.: 97.58 3rd Qu.: 782.5 3rd Qu.:73.7
## Max. :0.0009588 Max. :103.71 Max. :1043.0 Max. :73.7
## rela_slope_to_lowest rela_intercept_to_lowest rela_slope_to_lowest_cumsum
## Min. :0.9108 Min. :-823.49 Min. : 1.0
## 1st Qu.:1.0000 1st Qu.:-601.87 1st Qu.: 297.9
## Median :1.0000 Median :-101.68 Median : 735.0
## Mean :2.6677 Mean :-286.05 Mean : 990.1
## 3rd Qu.:5.7284 3rd Qu.: -76.81 3rd Qu.:1599.0
## Max. :5.7284 Max. : 0.00 Max. :2782.4
## rela_intercept_to_lowest_cumsum rela_slope_to_lowest_cumsum_invert
## Min. :-298346 Min. :0.0003594
## 1st Qu.:-151286 1st Qu.:0.0006254
## Median : -58624 Median :0.0013606
## Mean : -91892 Mean :0.0066321
## 3rd Qu.: -19563 3rd Qu.:0.0033568
## Max. : 0 Max. :1.0000000
## rela_intercept_to_lowest_cumsum_invert rela_x_intercept
## Min. : 0.00 Min. : 0.00
## 1st Qu.: 65.67 1st Qu.: 79.01
## Median : 79.76 Median : 98.16
## Mean : 78.57 Mean : 96.11
## 3rd Qu.: 94.61 3rd Qu.:114.36
## Max. :107.22 Max. :198.59
## opti_lowest_spline_knots tot_devi allocate_lowest opti_allocate
## Min. : 0 Min. :-22823 Min. :107.9 Min. : 0.00
## 1st Qu.: 3973 1st Qu.:-18849 1st Qu.:107.9 1st Qu.: 0.00
## Median : 13521 Median : -9301 Median :107.9 Median : 14.17
## Mean : 21151 Mean : -1672 Mean :107.9 Mean : 21.88
## 3rd Qu.: 31587 3rd Qu.: 8764 3rd Qu.:107.9 3rd Qu.: 33.13
## Max. :254218 Max. :231396 Max. :107.9 Max. :302.98
## allocate_total opti_exp_outcome
## Min. :22823 Min. :77.44
## 1st Qu.:22823 1st Qu.:78.36
## Median :22823 Median :80.19
## Mean :22823 Mean :79.45
## 3rd Qu.:22823 3rd Qu.:80.19
## Max. :22823 Max. :84.61## Warning: Unknown columns: `lowest_rank_alpha`, `lowest_rank_beta`## [1] ""
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxx rho:-25"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## S.country vil.id indi.id svymthRound
## Length:1043 Min. : 1.00 Min. : 2.0 Min. :24
## Class :character 1st Qu.: 7.00 1st Qu.: 298.5 1st Qu.:24
## Mode :character Median :12.00 Median : 627.0 Median :24
## Mean :12.52 Mean : 634.8 Mean :24
## 3rd Qu.:17.00 3rd Qu.: 939.5 3rd Qu.:24
## Max. :33.00 Max. :1349.0 Max. :24
## alpha alpha_log A A_log
## Min. :0.01014 Min. :0.006687 Min. :73.70 Min. :4.287
## 1st Qu.:0.01014 1st Qu.:0.006687 1st Qu.:77.73 1st Qu.:4.334
## Median :0.06018 Median :0.012090 Median :78.45 Median :4.343
## Mean :0.04108 Mean :0.010321 Mean :78.46 Mean :4.343
## 3rd Qu.:0.06018 3rd Qu.:0.012090 3rd Qu.:79.15 3rd Qu.:4.351
## Max. :0.06619 Max. :0.013940 Max. :84.61 Max. :4.417
## beta rank_val rank_idx lowest_rank_A
## Min. :0.0009588 Min. :107.3 Min. : 1.0 Min. :73.7
## 1st Qu.:0.0009588 1st Qu.:114.2 1st Qu.: 261.5 1st Qu.:73.7
## Median :0.0009588 Median :116.7 Median : 522.0 Median :73.7
## Mean :0.0009588 Mean :117.2 Mean : 522.0 Mean :73.7
## 3rd Qu.:0.0009588 3rd Qu.:120.5 3rd Qu.: 782.5 3rd Qu.:73.7
## Max. :0.0009588 Max. :126.1 Max. :1043.0 Max. :73.7
## rela_slope_to_lowest rela_intercept_to_lowest rela_slope_to_lowest_cumsum
## Min. :0.9125 Min. :-1055.06 Min. : 1.0
## 1st Qu.:1.0000 1st Qu.: -833.45 1st Qu.: 259.5
## Median :1.0000 Median : -99.67 Median : 513.6
## Mean :2.6023 Mean : -366.22 Mean : 834.7
## 3rd Qu.:5.5393 3rd Qu.: -75.96 3rd Qu.:1300.9
## Max. :5.5393 Max. : 0.00 Max. :2714.1
## rela_intercept_to_lowest_cumsum rela_slope_to_lowest_cumsum_invert
## Min. :-381972 Min. :0.0003684
## 1st Qu.:-146435 1st Qu.:0.0007687
## Median : -39659 Median :0.0019471
## Mean : -93578 Mean :0.0070255
## 3rd Qu.: -16917 3rd Qu.:0.0038538
## Max. : 0 Max. :1.0000000
## rela_intercept_to_lowest_cumsum_invert rela_x_intercept
## Min. : 0.00 Min. : 0.00
## 1st Qu.: 65.19 1st Qu.: 79.02
## Median : 77.22 Median :108.08
## Mean : 86.34 Mean :112.93
## 3rd Qu.:112.56 3rd Qu.:151.52
## Max. :140.73 Max. :215.41
## opti_lowest_spline_knots tot_devi allocate_lowest opti_allocate
## Min. : 0 Min. :-22823 Min. :120.6 Min. : 0.00
## 1st Qu.: 3588 1st Qu.:-19234 1st Qu.:120.6 1st Qu.: 0.00
## Median : 15848 Median : -6975 Median :120.6 Median : 13.87
## Mean : 29311 Mean : 6489 Mean :120.6 Mean : 21.88
## 3rd Qu.: 50677 3rd Qu.: 27855 3rd Qu.:120.6 3rd Qu.: 40.99
## Max. :202670 Max. :179848 Max. :120.6 Max. :121.45
## allocate_total opti_exp_outcome
## Min. :22823 Min. :75.60
## 1st Qu.:22823 1st Qu.:78.36
## Median :22823 Median :80.96
## Mean :22823 Mean :79.79
## 3rd Qu.:22823 3rd Qu.:80.96
## Max. :22823 Max. :84.61## Warning: Unknown columns: `lowest_rank_alpha`, `lowest_rank_beta`## [1] ""
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxx rho:-10"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## [1] "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
## S.country vil.id indi.id svymthRound
## Length:1043 Min. : 1.00 Min. : 2.0 Min. :24
## Class :character 1st Qu.: 7.00 1st Qu.: 298.5 1st Qu.:24
## Mode :character Median :12.00 Median : 627.0 Median :24
## Mean :12.52 Mean : 634.8 Mean :24
## 3rd Qu.:17.00 3rd Qu.: 939.5 3rd Qu.:24
## Max. :33.00 Max. :1349.0 Max. :24
## alpha alpha_log A A_log
## Min. :0.01014 Min. :0.006687 Min. :73.70 Min. :4.287
## 1st Qu.:0.01014 1st Qu.:0.006687 1st Qu.:77.73 1st Qu.:4.334
## Median :0.06018 Median :0.012090 Median :78.45 Median :4.343
## Mean :0.04108 Mean :0.010321 Mean :78.46 Mean :4.343
## 3rd Qu.:0.06018 3rd Qu.:0.012090 3rd Qu.:79.15 3rd Qu.:4.351
## Max. :0.06619 Max. :0.013940 Max. :84.61 Max. :4.417
## beta rank_val rank_idx lowest_rank_A
## Min. :0.0009588 Min. :179.0 Min. : 1.0 Min. :73.7
## 1st Qu.:0.0009588 1st Qu.:190.1 1st Qu.: 261.5 1st Qu.:73.7
## Median :0.0009588 Median :194.0 Median : 522.0 Median :73.7
## Mean :0.0009588 Mean :202.8 Mean : 522.0 Mean :73.7
## 3rd Qu.:0.0009588 3rd Qu.:220.6 3rd Qu.: 782.5 3rd Qu.:73.7
## Max. :0.0009588 Max. :227.1 Max. :1043.0 Max. :73.7
## rela_slope_to_lowest rela_intercept_to_lowest rela_slope_to_lowest_cumsum
## Min. :0.917 Min. :-1659.82 Min. : 1.0
## 1st Qu.:1.000 1st Qu.:-1438.21 1st Qu.: 257.5
## Median :1.000 Median : -94.39 Median : 505.2
## Mean :2.431 Mean : -575.93 Mean : 787.4
## 3rd Qu.:5.046 3rd Qu.: -73.43 3rd Qu.:1221.2
## Max. :5.046 Max. : 0.00 Max. :2535.6
## rela_intercept_to_lowest_cumsum rela_slope_to_lowest_cumsum_invert
## Min. :-600691 Min. :0.0003944
## 1st Qu.:-203379 1st Qu.:0.0008189
## Median : -37596 Median :0.0019792
## Mean :-129856 Mean :0.0070661
## 3rd Qu.: -16569 3rd Qu.:0.0038832
## Max. : 0 Max. :1.0000000
## rela_intercept_to_lowest_cumsum_invert rela_x_intercept
## Min. : 0.00 Min. : 0.00
## 1st Qu.: 64.34 1st Qu.: 76.34
## Median : 74.41 Median :102.59
## Mean :109.21 Mean :162.65
## 3rd Qu.:166.54 3rd Qu.:285.05
## Max. :236.91 Max. :328.97
## opti_lowest_spline_knots tot_devi allocate_lowest opti_allocate
## Min. : 0 Min. :-22823 Min. :119.1 Min. : 0.00
## 1st Qu.: 3091 1st Qu.:-19732 1st Qu.:119.1 1st Qu.: 0.00
## Median : 14234 Median : -8588 Median :119.1 Median : 15.17
## Mean : 63977 Mean : 41154 Mean :119.1 Mean : 21.88
## 3rd Qu.:144722 3rd Qu.:121899 3rd Qu.:119.1 3rd Qu.: 41.46
## Max. :233432 Max. :210609 Max. :119.1 Max. :119.13
## allocate_total opti_exp_outcome
## Min. :22823 Min. :74.37
## 1st Qu.:22823 1st Qu.:78.35
## Median :22823 Median :80.87
## Mean :22823 Mean :79.82
## 3rd Qu.:22823 3rd Qu.:80.87
## Max. :22823 Max. :84.61