Optimization problems are classified according to the types of values their decision variables can take. In continuous optimization, variables may take any real value within their allowed intervals. Examples include choosing a machine temperature, adjusting a material thickness, or determining a fluid flow rate. In discrete optimization, variables are restricted to separate values, often integers or binary values. Examples include deciding how many vehicles to purchase or whether to open a facility. Some models contain both continuous and discrete variables and are called mixed-integer optimization problems. The distinction matters because different mathematical structures require different algorithms. Continuous problems may be addressed using calculus, gradient methods, or nonlinear programming, while discrete problems may require combinatorial optimization, branch-and-bound, cutting planes, or specialized search methods.
Continuous and Discrete Optimization
Continuous optimization allows decision variables to take values from continuous numerical ranges, whereas discrete optimization restricts them to distinct values such as integers or binary choices.
Continuous: x ∈ ℝ
Integer: x ∈ ℤ
Binary: x ∈ {0, 1}
Mixed-integer: xᵢ ∈ ℝ for some variables and xⱼ ∈ ℤ for others
Explanation
Variables
ℝ: set of real numbers
ℤ: set of integers
x: decision variable
xᵢ: continuous variable
xⱼ: discrete variable
Example or application
The amount of electricity generated may be represented by a continuous variable, while the number of power stations to construct must be represented by an integer variable.
Definition
Continuous optimization is optimization over variables that can take values from continuous domains, while discrete optimization is optimization over variables restricted to discrete sets of allowable values.
Purpose
The purpose is to represent the true nature of decision choices and select solution methods appropriate to their mathematical domains.
Importance
Variable type strongly affects the difficulty and solution approach. Treating an integer decision as continuous may produce an impractical fractional result.
Learning Objectives
Define continuous and discrete optimization.
Distinguish real-valued, integer, and binary variables.
Explain mixed-integer optimization.
Identify practical examples of each type.
Understand why variable domains influence algorithm selection.
Key Concepts
Continuous variable
Integer variable
Binary variable
Discrete decision
Mixed-integer optimization
Combinatorial optimization
Working Principle
The variable domain determines which values are permitted. A continuous method searches over real-valued choices, while a discrete method must respect restrictions such as integer or binary decisions.
Process
Identify each decision that must be represented.
Determine whether fractional values are meaningful.
Assign a continuous, integer, or binary domain.
Formulate the objective and constraints.
Choose an algorithm compatible with the variable domains.
Verify that the final solution satisfies the required domains.
Applications
Continuous control and engineering design
Facility-location decisions
Vehicle and equipment selection
Workforce assignment
Power generation and capacity planning
Worked Example
A company chooses how many machines to purchase. Let x be the number of machines. Since purchasing 2.5 machines is not meaningful, the model must require x ∈ ℤ and x ≥ 0. If each machine costs $4,000 and the budget is $20,000, the constraint is 4000x ≤ 20000. The feasible integer values are 0, 1, 2, 3, 4, and 5.
Common Misconceptions
Discrete optimization is not limited to binary variables; integer variables may have many possible values.
A continuous solution cannot automatically be rounded without checking feasibility and optimality.
Mixed-integer models contain both continuous and discrete variables.
Continuous optimization is not necessarily easy; nonlinear continuous problems can also be difficult.
Practice Questions
Define continuous and discrete optimization.
Classify the number of vehicles purchased as continuous or discrete.
What is a binary decision variable?
Explain mixed-integer optimization with an example.
Why can rounding a continuous solution produce an infeasible decision?
Key Takeaway
The domain of decision variables determines whether an optimization problem is continuous, discrete, or mixed-integer and strongly influences how it should be solved.
Prerequisites
Decision Variables and Parameters
Objective Functions
Basic number systems
Reference
Wolsey, L. A. Integer Programming.
Bertsimas, D., and Tsitsiklis, J. N. Introduction to Linear Optimization.
Related Topics
Deterministic and Stochastic Optimization
Constrained and Unconstrained Optimization
Optimization Algorithms and Solution Methods
Optimization Problem Formulation
Sources and verification
Sources are shown as provided in the project data. Not every source has yet been independently verified. Please report errors or missing citations.
- Source verification pending; check an authoritative textbook, university, standards body, or peer-reviewed source.
Suggest a correction or source · More in Optimization & Operations Research