Appendix C — Quick Reference

A compact reference to every exported function in the PRA package. Arguments marked * are required; all others have defaults.

C.1 Monte Carlo Simulation

Function	Key Arguments	Returns	Chapter
`mcs(num_sims, task_dists, cor_mat)`	`num_sims` number of simulations; `task_dists` list of distribution specs; optional correlation matrix	S3 object of class `mcs` with `$simulations` matrix, `$total_distribution`, percentiles	Chapter 2
`print.mcs(x)`	`x` an `mcs` object	Console summary of P50/P80/P95 and task means	Chapter 2

Distribution spec format (used by mcs(), sensitivity(), smm()):

list(type = "normal",     mean = 10, sd = 2)
list(type = "triangular", a = 5, b = 10, c = 15)   # a=min, b=mode, c=max
list(type = "uniform",    min = 8, max = 12)

C.2 Second Moment Method

Function	Key Arguments	Returns	Chapter
`smm(mean, var, cor_mat)`	Numeric vectors of task means and variances; optional correlation matrix	S3 object of class `smm` with `$total_mean`, `$total_var`, `$total_sd`	Chapter 3
`print.smm(x)`	`x` an `smm` object	Console summary of total mean, SD, and 95% CI	Chapter 3

C.3 Sensitivity Analysis

Function	Key Arguments	Returns	Chapter
`sensitivity(task_dists*, cor_mat)`	Same distribution specs as `mcs()`; optional correlation matrix	Named numeric vector of sensitivity indices (one per task)	Chapter 4

C.4 Contingency & Correlation

Function	Key Arguments	Returns	Chapter
`contingency(mcs_result*, phigh, pbase)`	`mcs_result` from `mcs()`; `phigh` target percentile (default 0.95); `pbase` base percentile (default 0.50)	Numeric contingency reserve (phigh quantile minus pbase quantile)	Chapter 2
`cor_matrix(num_samples, num_vars, dists*)`	`dists` list of sampling functions `function(n) ...`; `num_samples` (default 100); `num_vars` (default 5)	Numeric correlation matrix	Chapter 4

C.5 Earned Value Management

All EVM functions accept numeric scalars or vectors as noted.

Function	Key Arguments	Returns	Chapter
`pv(bac, schedule, time_period*)`	Budget at completion; cumulative schedule proportions; current period (1-based)	Planned Value	Chapter 5
`ev(bac, actual_per_complete)`	Budget at completion; fraction complete (0–1)	Earned Value	Chapter 5
`ac(actual_costs, time_period, cumulative)`	Vector of actual costs per period; current period; whether `actual_costs` is already cumulative (default `TRUE`)	Actual Cost	Chapter 5
`sv(ev, pv)`	Earned Value; Planned Value	Schedule Variance (EV − PV)	Chapter 5
`cv(ev, ac)`	Earned Value; Actual Cost	Cost Variance (EV − AC)	Chapter 5
`spi(ev, pv)`	Earned Value; Planned Value	Schedule Performance Index (EV / PV)	Chapter 5
`cpi(ev, ac)`	Earned Value; Actual Cost	Cost Performance Index (EV / AC)	Chapter 5
`eac(bac*, method, cpi, ac, ev, spi)`	BAC; `method` = “typical” / “atypical” / “combined” (default “typical”); `cpi` required for typical; `ac`/`ev` required for atypical; `cpi`/`ac`/`ev`/`spi` required for combined	Estimate at Completion	Chapter 5
`etc(bac, ev, cpi)`	Budget at Completion; Earned Value; Cost Performance Index	Estimate to Complete	Chapter 5
`tcpi(bac, ev, ac*, target, eac)`	BAC, EV, AC; `target` = “bac” or “eac” (default “bac”); `eac` required when `target = "eac"`	To-Complete Performance Index	Chapter 5
`vac(bac, eac)`	Budget at Completion; Estimate at Completion	Variance at Completion (BAC − EAC)	Chapter 5

C.6 Bayesian Risk

Function	Key Arguments	Returns	Chapter
`risk_prob(causes, given, not_given*)`	Vectors of prior probabilities, likelihoods if cause present/absent	Prior risk probability (scalar)	Chapter 6
`risk_post_prob(causes, given, not_given, observed)`	Same as above plus `observed` vector (1 = observed, 0 = not, NA = unknown)	Posterior risk probability after observations	Chapter 6

C.7 Learning Curves

Function	Key Arguments	Returns	Chapter
`fit_sigmoidal(data, x_col, y_col, model_type)`	`data` a data frame; `x_col`/`y_col` column names as strings; `model_type` = “logistic” / “gompertz” / “pearl”	Fitted `nls` model object	Chapter 7
`predict_sigmoidal(fit, x_range, model_type*, conf_level)`	Object from `fit_sigmoidal()`; new x values; model type; optional confidence level	Data frame of predictions (with confidence bounds if `conf_level` set)	Chapter 7
`plot_sigmoidal(fit, data, x_col, y_col, model_type*, conf_level)`	Object from `fit_sigmoidal()`; original data frame and column names; model type; optional confidence level	Plot of fitted curve with confidence band	Chapter 7

C.8 Design Structure Matrices

Function	Key Arguments	Returns	Chapter
`parent_dsm(S*)`	Resource-Task matrix S (resources × tasks, binary)	S3 object of class `dsm` (`$type` is `"parent"`); `$matrix` is tasks × tasks	Chapter 9
`grandparent_dsm(S, R)`	Resource-Task matrix S; Risk-Resource matrix R (risks × resources, binary)	S3 object of class `dsm` (`$type` is `"grandparent"`); `$matrix` is tasks × tasks	Chapter 9

Both constructors return an object of class dsm, which has print(), summary() and plot() methods.

C.9 Probabilistic Networks

Function	Key Arguments	Returns	Chapter
`prob_net(nodes, links, distributions*)`	Data frames of nodes (id, label, group) and links (source, target, value); named list of distributions	Graph object (list of nodes, links, distributions)	Chapter 8
`prob_net_sim(graph*, num_samples)`	Graph from `prob_net()`; number of simulations (default 10 000)	Named list of sample vectors, one per node	Chapter 8
`prob_net_learn(graph, observations, num_samples)`	Graph; named list of clamped observations e.g. `list(B = "No")`; sample count	Named list of sample vectors with upstream nodes clamped	Chapter 8
`prob_net_update(graph*, remove_links, update_distributions)`	Graph; data frame of edges to remove; named list of replacement distributions	Modified graph object	Chapter 8

Node distribution types (used in distributions list):

list(type = "discrete",     values = c(1, 0), probs = c(0.7, 0.3))
list(type = "normal",       mean = 50000, sd = 10000)
list(type = "conditional",  condition = "A",
     true_dist  = list(type = "normal", mean = 80000, sd = 20000),
     false_dist = list(type = "normal", mean = 50000, sd = 10000))
list(type = "aggregate",    nodes = c("F", "G", "H"))

C.10 Agentic Framework

Function	Key Arguments	Returns	Chapter
`pra_tools()`	None	List of `ellmer` tool definitions wrapping the PRA functions	Chapter 12
`pra_mcp_server()`	None	Starts MCP server; register with `claude mcp add`	Chapter 12

# Quick Reference {#sec-ref} A compact reference to every exported function in the `PRA` package. Arguments marked `*` are required; all others have defaults. ## Monte Carlo Simulation | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `mcs(num_sims*, task_dists*, cor_mat)` | `num_sims` number of simulations; `task_dists` list of distribution specs; optional correlation matrix | S3 object of class `mcs` with `$simulations` matrix, `$total_distribution`, percentiles | @sec-mcs | | `print.mcs(x)` | `x` an `mcs` object | Console summary of P50/P80/P95 and task means | @sec-mcs | **Distribution spec format** (used by `mcs()`, `sensitivity()`, `smm()`): ```r list(type = "normal", mean = 10, sd = 2) list(type = "triangular", a = 5, b = 10, c = 15) # a=min, b=mode, c=max list(type = "uniform", min = 8, max = 12) ``` --- ## Second Moment Method | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `smm(mean*, var*, cor_mat)` | Numeric vectors of task means and variances; optional correlation matrix | S3 object of class `smm` with `$total_mean`, `$total_var`, `$total_sd` | @sec-smm | | `print.smm(x)` | `x` an `smm` object | Console summary of total mean, SD, and 95% CI | @sec-smm | --- ## Sensitivity Analysis | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `sensitivity(task_dists*, cor_mat)` | Same distribution specs as `mcs()`; optional correlation matrix | Named numeric vector of sensitivity indices (one per task) | @sec-sensitivity | --- ## Contingency & Correlation | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `contingency(mcs_result*, phigh, pbase)` | `mcs_result` from `mcs()`; `phigh` target percentile (default 0.95); `pbase` base percentile (default 0.50) | Numeric contingency reserve (phigh quantile minus pbase quantile) | @sec-mcs | | `cor_matrix(num_samples, num_vars, dists*)` | `dists` list of sampling functions `function(n) ...`; `num_samples` (default 100); `num_vars` (default 5) | Numeric correlation matrix | @sec-sensitivity | --- ## Earned Value Management All EVM functions accept numeric scalars or vectors as noted. | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `pv(bac*, schedule*, time_period*)` | Budget at completion; cumulative schedule proportions; current period (1-based) | Planned Value | @sec-evm | | `ev(bac*, actual_per_complete*)` | Budget at completion; fraction complete (0–1) | Earned Value | @sec-evm | | `ac(actual_costs*, time_period*, cumulative)` | Vector of actual costs per period; current period; whether `actual_costs` is already cumulative (default `TRUE`) | Actual Cost | @sec-evm | | `sv(ev*, pv*)` | Earned Value; Planned Value | Schedule Variance (EV − PV) | @sec-evm | | `cv(ev*, ac*)` | Earned Value; Actual Cost | Cost Variance (EV − AC) | @sec-evm | | `spi(ev*, pv*)` | Earned Value; Planned Value | Schedule Performance Index (EV / PV) | @sec-evm | | `cpi(ev*, ac*)` | Earned Value; Actual Cost | Cost Performance Index (EV / AC) | @sec-evm | | `eac(bac*, method, cpi, ac, ev, spi)` | BAC; `method` = "typical" / "atypical" / "combined" (default "typical"); `cpi` required for typical; `ac`/`ev` required for atypical; `cpi`/`ac`/`ev`/`spi` required for combined | Estimate at Completion | @sec-evm | | `etc(bac*, ev*, cpi)` | Budget at Completion; Earned Value; Cost Performance Index | Estimate to Complete | @sec-evm | | `tcpi(bac*, ev*, ac*, target, eac)` | BAC, EV, AC; `target` = "bac" or "eac" (default "bac"); `eac` required when `target = "eac"` | To-Complete Performance Index | @sec-evm | | `vac(bac*, eac*)` | Budget at Completion; Estimate at Completion | Variance at Completion (BAC − EAC) | @sec-evm | --- ## Bayesian Risk | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `risk_prob(causes*, given*, not_given*)` | Vectors of prior probabilities, likelihoods if cause present/absent | Prior risk probability (scalar) | @sec-bayes | | `risk_post_prob(causes*, given*, not_given*, observed*)` | Same as above plus `observed` vector (1 = observed, 0 = not, NA = unknown) | Posterior risk probability after observations | @sec-bayes | --- ## Learning Curves | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `fit_sigmoidal(data*, x_col*, y_col*, model_type*)` | `data` a data frame; `x_col`/`y_col` column names as strings; `model_type` = "logistic" / "gompertz" / "pearl" | Fitted `nls` model object | @sec-sigmoidal | | `predict_sigmoidal(fit*, x_range*, model_type*, conf_level)` | Object from `fit_sigmoidal()`; new x values; model type; optional confidence level | Data frame of predictions (with confidence bounds if `conf_level` set) | @sec-sigmoidal | | `plot_sigmoidal(fit*, data*, x_col*, y_col*, model_type*, conf_level)` | Object from `fit_sigmoidal()`; original data frame and column names; model type; optional confidence level | Plot of fitted curve with confidence band | @sec-sigmoidal | --- ## Design Structure Matrices | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `parent_dsm(S*)` | Resource-Task matrix S (resources × tasks, binary) | S3 object of class `dsm` (`$type` is `"parent"`); `$matrix` is tasks × tasks | @sec-dsm | | `grandparent_dsm(S*, R*)` | Resource-Task matrix S; Risk-Resource matrix R (risks × resources, binary) | S3 object of class `dsm` (`$type` is `"grandparent"`); `$matrix` is tasks × tasks | @sec-dsm | Both constructors return an object of class `dsm`, which has `print()`, `summary()` and `plot()` methods. --- ## Probabilistic Networks | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `prob_net(nodes*, links*, distributions*)` | Data frames of nodes (id, label, group) and links (source, target, value); named list of distributions | Graph object (list of nodes, links, distributions) | @sec-network | | `prob_net_sim(graph*, num_samples)` | Graph from `prob_net()`; number of simulations (default 10 000) | Named list of sample vectors, one per node | @sec-network | | `prob_net_learn(graph*, observations*, num_samples)` | Graph; named list of clamped observations e.g. `list(B = "No")`; sample count | Named list of sample vectors with upstream nodes clamped | @sec-network | | `prob_net_update(graph*, remove_links, update_distributions)` | Graph; data frame of edges to remove; named list of replacement distributions | Modified graph object | @sec-network | **Node distribution types** (used in `distributions` list): ```r list(type = "discrete", values = c(1, 0), probs = c(0.7, 0.3)) list(type = "normal", mean = 50000, sd = 10000) list(type = "conditional", condition = "A", true_dist = list(type = "normal", mean = 80000, sd = 20000), false_dist = list(type = "normal", mean = 50000, sd = 10000)) list(type = "aggregate", nodes = c("F", "G", "H")) ``` --- ## Agentic Framework | Function | Key Arguments | Returns | Chapter | |----------|--------------|---------|---------| | `pra_tools()` | None | List of `ellmer` tool definitions wrapping the PRA functions | @sec-agent | | `pra_mcp_server()` | None | Starts MCP server; register with `claude mcp add` | @sec-agent |