Economics is all about choices. A family decides how to spend its income. A company decides how many people to hire. A government decides how to allocate its budget. At the heart of all these decisions is a single goal: getting the best possible outcome with limited resources. This process of finding the “best” is called optimization.

To frame these choices mathematically, we use two key components. First is the objective function, which represents the goal we want to achieve. This could be maximizing profit, happiness (what economists call utility), or social welfare. It could also be minimizing costs, waste, or risk. Second, we have constraints, which are the limitations or rules we must follow. These are things like budgets, time, physical resources, or regulations.

For example, a simple consumer problem can be framed like this:

• Objective: Maximize happiness from buying goods.
• Constraint: You can't spend more money than you have.

Putting this into mathematical language is the first step toward solving the problem.

Many economic decisions are about the here and now. Static optimization deals with these single-period decisions. It's like taking a snapshot in time and finding the best possible choice within that frame. Let's return to our consumer, who is choosing between two goods: apples ($x$) and bananas ($y$).

Their objective is to maximize their utility function, $U(x, y)$, which measures their satisfaction. Their constraint is their budget. If apples cost $p_x$ each and bananas cost $p_y$, and their total income is $I$, they cannot spend more than $I$. The problem is:

How do we solve this? A powerful technique is the method of Lagrange multipliers. We create a new function, called the Lagrangian, that combines the objective function and the constraint into one.

Here, $\lambda$ (lambda) is the Lagrange multiplier. It has a neat economic interpretation: it represents the marginal utility of income, or how much your happiness would increase if you had one more dollar to spend.

The key to finding the optimal solution is to find the point where the slope of the indifference curve (representing the consumer's willingness to trade one good for another) is exactly equal to the slope of the budget constraint (representing the market's trade-off rate). This is the point of tangency, where the consumer gets the most happiness for their money.

Finding that tangency point requires using calculus. By taking the partial derivatives of the Lagrangian function with respect to $x$, $y$, and $\lambda$ and setting them to zero, we get what are known as the first-order conditions. These are necessary conditions for an optimum.

For our consumer, these conditions boil down to a simple, intuitive rule:

This equation says a rational consumer will allocate their spending so that the last dollar spent on apples gives them the same amount of extra happiness as the last dollar spent on bananas. If it didn't, they could rearrange their spending to become happier.

First-order conditions identify potential solutions, but they don't distinguish between a maximum (a hilltop), a minimum (a valley floor), or a saddle point. To confirm we've found a maximum, we need to check the second-order conditions, which relate to the curvature of the function. For our consumer problem, the convex shape of the indifference curves ensures we've found a point of maximum utility.

But what about decisions that span across time? People don't just decide what to buy today; they decide how much to save for retirement, how much to invest in education, or when to harvest a forest. These problems require dynamic optimization.

Instead of a single objective function, dynamic optimization involves maximizing a sum of outcomes over a period of time. A common example is deciding how much to consume and save over one's lifetime. The goal is to maximize total lifetime utility. The choice you make today (how much to save) affects the resources you have tomorrow, creating a link across time periods.

Let's say a person lives for two periods, today (period 1) and tomorrow (period 2). Their objective is to maximize the sum of their utility from consumption ($C$) in both periods. But future happiness is often seen as less valuable than present happiness, so we discount future utility by a factor $\beta$ (beta).

The constraint is that total consumption across both periods cannot exceed total income. If they earn $I_1$ and $I_2$ and can save at an interest rate $r$, the constraint links the two periods:

This is known as the intertemporal budget constraint. Techniques like dynamic programming or optimal control theory are used to solve these problems. The solution yields an optimal path of consumption and savings over time, balancing the desire for satisfaction today against the need for resources tomorrow.

Whether looking at a single moment or across a lifetime, optimization provides the framework for understanding economic decision-making as a purposeful act of balancing goals against limitations.