Feasibility and Solution Assessment Questions

import numpy as np

def baseline_contingency(risks):
    # risks: list of dicts {'name','p','impact'}
    breakdown = {r['name']: r['p']*r['impact'] for r in risks}
    total = sum(breakdown.values())
    return total, breakdown

def correlated_contingency(risks, corr_matrix=None, sims=10000, percentile=90, seed=0):
    np.random.seed(seed)
    n = len(risks)
    p = np.array([r['p'] for r in risks])
    impacts = np.array([r['impact'] for r in risks])
    # default independence
    if corr_matrix is None:
        corr_matrix = np.eye(n)
    # convert probabilities to normal thresholds (Gaussian copula)
    from scipy.stats import norm
    thresholds = norm.ppf(p)
    L = np.linalg.cholesky(corr_matrix)
    samples = np.zeros(sims)
    for i in range(sims):
        z = np.random.normal(size=n)
        correlated = L @ z
        events = (correlated < thresholds).astype(float)  # 1 = risk occurs
        samples[i] = np.sum(events * impacts)
    recommended = np.percentile(samples, percentile)
    return recommended, samples

# Example usage:
# risks = [{'name':'A','p':0.2,'impact':10000}, ...]
# baseline_contingency(risks)
# correlated_contingency(risks, corr_matrix, sims=50000)