Lý Thuyết

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 3

1.1.

Stochastic linear programs (SLP) are linear programs (i.e. its objective function is linear) in which

some problem data may be considered uncertain.

A stochastic linear program (SLP) is a type of mathematical optimization problem that involves linear
objective and constraints, but also some uncertain or random parameters. SLP can be used to model
decision making under uncertainty, such as in finance, energy, or logistics.

According to part 4.1 of the current page, a stochastic linear program (SLP) is:

A type of mathematical optimization problem that involves linear objective and constraints, but with
some uncertain or random parameters.

A way to deal with randomness in problem formulation and solution, by using probabilistic or scenario-
based approaches.

A tool for decision making under uncertainty in various application areas, such as finance, energy,
telecommunication, etc.

1.2

Two-stage stochastic linear programming (2-SLP) is a type of optimization problem that involves making
decisions under uncertainty. In 2-SLP, there are two stages of decision making: the first stage, where
some decisions are made before the uncertainty is revealed, and the second stage, where some
corrective actions or recourse are taken after the uncertainty is realized. The objective is to minimize the
expected total cost of the first-stage and second-stage decisions, subject to some constraints.

2-SLP can be formulated as follows:

The first-stage decision variables are denoted by x, and the objective function is c^T x, where c is a vector
of coefficients.

The second-stage decision variables are denoted by y, and the objective function is q y, where q is a
vector of coefficients. The second-stage decision variables depend on the first-stage decision variables
and the realization of the random vector w, which represents the uncertainty.

The first-stage constraints are A x = b, where A is a matrix and b is a vector.

The second-stage constraints are T(w) x + W y = h(w), where T, W, and h are functions of the random
vector w.

1.2.1

Simple Form:

The simple form of the 2-SLP recourse model focuses on minimizing the expected total cost. It considers
both the cost of stage 1 decisions and the expected cost of stage 2 recourse actions.

1.2.2

Canonical Form:
The canonical form of the 2-SLP recourse model explicitly includes the uncertainty in the model
formulation. It uses decision variables and constraints for both stage 1 decisions and stage 2 recourse
actions.

1.3

A minimum cost maximum flow algorithm (MCMF) is a type of graph algorithm that finds the cheapest
way to send a certain amount of flow through a network. It is a fundamental problem in optimization
and has applications in various fields, including:

 Transportation networks
 Communication networks
 Supply chain management
 Project scheduling

How it works:

Input: A network is represented by a graph with a source node, a sink node, and edges representing
connections between nodes. Each edge has a capacity (maximum amount of flow it can carry) and a cost
(per unit of flow).

Goal: Find a flow f that satisfies the following conditions:

Flow conservation: The amount of flow entering a node equals the amount leaving it (except for the
source and sink nodes).

Capacity constraints: The flow on each edge cannot exceed its capacity.

Maximum flow: The total flow from the source to the sink is maximized.

Minimum cost: The total cost of the flow is minimized.

1.3.1

Ford-Fulkerson is a fundamental algorithm for finding the maximum flow in a network. It is a greedy
algorithm that starts with an empty flow and iteratively increases the flow by finding augmenting paths
(paths from the source to the sink that have available capacity).

There are several variations of the Ford-Fulkerson algorithm, each with its own advantages and
disadvantages. Here are some of the most common variants:

Ford-Fulkerson with Edmonds-Karp: This variant uses a more efficient way to find augmenting paths,
resulting in better time complexity.

Ford-Fulkerson with Dinic's algorithm: This variant further improves the efficiency of finding augmenting
paths by utilizing layered graphs.

These variants all share the basic principles of the Ford-Fulkerson algorithm and are primarily focused on
finding the maximum flow.

1.3.2
The Cycle Cancelling Algorithm is a different approach to finding the minimum cost maximum flow in a
network. It works by iteratively cancelling cycles in the residual graph (a modified version of the original
network) that have negative cost. By cancelling these cycles, the algorithm increases the flow while
reducing the overall cost.

Some of the key features of the Cycle Cancelling Algorithm:

 It can handle negative costs in the network.


 It can be more efficient than Ford-Fulkerson based methods for certain problem instances.
 It can be more challenging to understand and implement.

You might also like