Risk management

ORDECSYS designs quantitative models of risk management based on Dynamic Programming, or Stochastic/Robust Programming paradigms. These models deal in particular with the valuation of flexibility (options) in procurement contracts, the design of hedging policies in production planning and optimal investment policies under uncertainty.

Stochastic programming

Stochastic programming is an approach to perform an optimization over an event tree that represents different possible random scenarios. The method is particularly relevant when one assesses decisions that may have some irreversible effects.

Robust programming

Robust programming is a technique to assess decisions under uncertainty which is based on worst case analysis instead of a probability description of risk. The method is particularly useful for design problems where some key parameters are uncertain.

Dynamic programming

Dynamic programming is the standard approach to solve dynamic optimization problems under uncertainty, also known as 'stochastic control problems'.

Monte Carlo methods

A promising approach for solving large scale risk management problems consists in coupling a simulator generating random scenarios with an optimization technique permitting the identification of efficient decision rules.