imperial-institute

Stochastic Programming

Stochastic programming in the field of mathematical optimisation, it is a framework for modelling problems of optimisation that involves uncertainty. Whereas deterministic problems of optimisation are formulated parameters that are known, real world problems almost invariably encompass some unknown parameters. When the parameters are easily identified within certain bounds, an approach to handling such problems is known as robust optimisation.

Here the aim is to find a solution that is optimal in some sense and feasible for all such kind of data. Stochastic programming models are the same in style but undermine the fact that the probability distribution that governs the data can be estimated and are easily identified/ known. The aim here is to find a feasible policy for all the possible instances of data and maximises some functions’ expectations of the random variables and the decisions. Primarily, such models are solved numerically and analytically, analysed and formulated to provide a choice make with useful information.

Applications of stochastic programming


The applications of stochastic programming encompass two parts. The first part consist of stochastic programming systems that aid in analysing stochastic data. All the codes of these programs have been extensively developed and tested and will appear to developers and researchers who want to make models without implementation cost and extensive programming. These codes are the Synopsis of the best available systems, with the need that they are ready to go, accessible to the public and user-friendly. The second part is the importance of stochastic programming in the supply chain, electricity, financial modelling, telecommunication, pollution and environmental control and scheduling. These comprise the complete collection of the real application using available stochastic programming in the literature.

Several practical decision problems involve uncertainty. Stochastic programming involves the study of procedures for making decisions under uncertainty over time. This uncertainty can be the model itself or the in the model parameters. Parameters may be uncertain due to unobservable and future events, lack of reliable data and measurement errors. The events’ uncertainty, details of the problem constraint and structures and the payoff/risk of decisions are modelled in an optimisation framework.

Difficulties of stochastic programming


The primary problem of stochastic programming is the complexity involved in the entire procedure. The whole process is very complicated regardless whether it is manually or by the use of stochastic programs as they are a lot of calculations and data entry respectively. The method is also time consuming and tedious. Data entry into the stochastic programs is very tedious and requires full attention as a single mistake may result to an error in the entire.