Markov & state-space control
Choose a policy over evolving states when today's action changes tomorrow's options.
10 of 388 tools.
Estimate dynamic execution factor
Extract a direction-aligned latent execution factor from aggregate metric vectors and forecast its level and velocity with a likelihood-tuned local-linear-trend state-space model.
Estimate platform network option value
Value when to activate a shared platform under endogenous network adoption with an exact finite-horizon Markov dynamic program; optimize the invest/wait policy by observed adopter state, compare it with every fixed launch date and never investing, and reconcile option value, investment timing, and adoption quantiles.
Fit team behavior regime HMM
Learn persistent privacy-safe team operating regimes and transitions with a Gaussian hidden Markov model.
Forecast agentic action loss
Forecast expected and tail operational AI-agent loss with tenant-local empirical-Bayes absorbing Markov chains across execution, deviation, containment, recovery, completion and loss, preserving shared control-failure regimes and unresolved-chain mass.
Forecast alert fatigue and missed risk loss
Forecast alert storms, duplicate notifications, aggregate attention-state saturation, missed material conditions, interruption cost and financial VaR/CVaR with a Markov-modulated Gamma-Poisson and compound log-normal model.
Forecast recurring revenue regimes
Fit a diagonal-Gaussian hidden Markov model to consecutive organic ARR growth, GRR, gross margin and cash-burn intensity, order latent regimes by growth rather than arbitrary labels, test effective regime support and improvement over a single-state model, then simulate ARR and unrestricted cash through fitted transitions, within-regime variation and coherent common market scenarios.
Optimize fundraising attention policy
Allocate the current fundraising attention epoch with age-aware controlled Markov arm values and an exact multiple-choice capacity knapsack: compare action versus passive continuation through later stages, price founder distraction and action cost, expose a dynamic attention index, and disclose that independently relaxed future capacity is not a globally certified multi-period schedule.
Optimize multi period calibration maintenance
Optimize a finite-horizon analytics maintenance schedule by propagating each function's healthy/degraded Markov belief under passive operation or recalibration, valuing healthy decisions and uncalibrated loss, enforcing period cash and specialist-capacity constraints, and disclosing exact state enumeration versus deterministic beam search.
Solve distributionally robust markov policy
Solve a discounted Markov policy against simultaneous L1 transition-confidence sets derived from empirical state-action counts.
Solve entropic risk sensitive markov policy
Solve a finite-horizon Markov policy under exponential downside utility and compare it with the risk-neutral policy using paired Monte Carlo lower-tail CVaR.