Mathematical optimization - Short Note

Mathematical optimization is the rigorous discipline of identifying the best possible solution—minimizing costs or maximizing outputs—from a set of available alternatives based on a defined objective, subject to specific limits. Variables, constants, and equations form the mathematical foundations that allow us to model these real-world problems. [1, 2]

1. Variables: The Control Inputs

In optimization, variables represent the physical quantities or parameters you can control, change, or determine. They define the space of possible solutions. [1, 2, 3, 4, 5]

• Decision Variables (\(x\)): The core unknown values we are trying to find. They can be singular (\(x\)) or represented as a vector (\(x_1, x_2, \ldots, x_n\)).
• Continuous Variables: Can take any value within a range (e.g., length, time, or temperature).
• Discrete Variables: Can only take specific, distinct values (e.g., number of machines, number of items produced). [1, 2, 3, 4, 5]

2. Constants: The Rules and Limits

Constants are fixed numerical values or parameters that do not change during the optimization process. They shape the environment in which your variables operate. [1]

• Fixed Data Parameters: Known values like material densities, gravitational pull, or inflation rates.
• Bounds: Upper and lower limits applied to variables (e.g., \(x \ge 0\)). [1, 2, 3, 4, 5]

3. Equations: The Objective and Constraints

Equations mathematically bind the variables and constants together to represent reality. Every standard optimization model relies on two primary types of equations: [1, 2, 3]

A. The Objective Function

This is the specific equation you are trying to maximize (e.g., profit, efficiency) or minimize (e.g., cost, error). [1, 2]

• Example: \(Minimize \ f(x) = 2x_{1}^{2} + 5x_{2}\)

B. Constraint Functions

These are equations (or inequalities) that place physical, logical, or theoretical limits on your variables. They restrict the "feasible region" of your solution. [1, 2, 3, 4, 5]

• Equality Constraints (\(h(x) = 0\)): Fixed requirements (e.g., total volume must exactly equal \(50 \ m^3\)).
• Inequality Constraints (\(g(x) \le 0\)): Limits that cannot be exceeded (e.g., working hours \(\le 40\) hours). [1, 2, 3, 4, 5]

A Practical Example

Imagine you are trying to build the most spacious rectangular fenced yard with a fixed amount of fencing material.

• Variables: \(x\) (width), \(y\) (length)
• Constant: \(500\) meters (total fencing material available)
• Objective Equation: Maximize the Area \(A = x \times y\)
• Constraint Equation: \(2x + 2y = 500\) [1, 2, 3]

To solve this, you would use algebra or calculus to substitute the constraint into the objective equation so it relies on only one variable, then find where the rate of change reaches zero. [1, 2]

Would you like to explore how to apply these concepts to Linear Programming, or would you prefer a deep dive into Calculus-based methods (like taking derivatives) to find optimal points?