What Is Optimization?

What is optimization?

Optimization is the essence of management and can be a fundamental factor affecting a company’s profitability, efficiency and product quality. We see companies that are developing well and offer products and services of excellent quality. Can it be assumed that the source of these entities’ success lies precisely in optimization? Nothing could be further from the truth. Companies throughout the world owe their development to a combination of the proper accumulation of generated capital, innovation and a market situation favorable to the given entity. Companies reinvest the profit they generate in modern means of production and new technologies, and thereby increase their efficiency. Unfortunately, this is not optimization, but accumulation.

Optimization is generally understood to be an action aimed at improving existing conditions, improving safety or reducing costs.

Optimization is often referred to as cost cuts or restructuring of enterprises to improve their efficiency.

Unfortunately, this common definition has little to do with real optimization. The optimization process involves taking action to achieve the desired marginal value.

Optimization is always about reaching the maximum level. Optimization methods didn't come into being until the 1960s.

We're talking, of course, about complex manufacturing or service processes that have many levels of constraints and limitations, and the result was the work of mathematics professor George Dantzig, who proposed ways to calculate extreme values for complex processes using graphical methods.

W-MOSZCZYNSKI-2021-7-18

What is and what is not optimization?

To understand what the definition of optimization means, we'll use a simple example: a butcher's company has to deliver its cutting-edge products to many wholesalers.

So the owner bought another, extra, larger delivery car, which could not be called an optimization, and the shop owner was convinced that his cars were being used to their full potential, but he didn't know how much of the maximum volume of goods he could carry on existing routes and driving time constraints.

But you can't blame him for his decision, he did what he thought was the wisest thing to do, and in this situation, there are a number of techniques that would maximize the efficiency of cars, like the transportation issue that sets the best routes, the rail issue that minimizes stops.

You can use methods that maximize the use of vehicle cargo space, and you can also use mathematical algorithms to calculate whether it's more costly to dispose of your fleet in favor of the cars you send to your customers.

If we define a process as operating under certain conditions, then optimization is finding and then achieving extreme values within that process.

If delivery vehicles are forced to cross a bridge, optimal routes can be determined for them. Optimization concerns a defined process. The appearance of a new bridge changes that process, and the determination of optimal routes must begin again. If one of the machines used to pack meat proves to be a „bottleneck”, replacing it with a new, more efficient machine is not optimization, but an investment. Paradoxically, a plant may record a decline in profit because purchasing the machine brought new, higher depreciation costs. Perhaps it would have been worthwhile first to run an operations-research algorithm indicating which product assortment leads to profit maximization.

It is also possible to analyse whether increasing productivity is a viable phenomenon, there are methods for determining optimal storage conditions or optimal downtime that lead to a significant reduction in operating costs.

The purchase of a new machine certainly solved the problems of the time, but it did not have to increase the efficiency of the plant or improve its profitability. Moreover, the advent of a new machine changed the process, which affected the existing solutions.

How to use optimization

Optimization is a difficult process, and it's not accessible to the average manager or analyst, and to apply it, you have to know a number of mathematical optimization methods, and you have to have programming skills to be able to use special libraries dedicated to optimizing libraries.

All of this is difficult, and it's not indicative that there's going to be some kind of application that's going to help with the management process, and interestingly, most of today's production and logistics management systems aren't equipped with optimization algorithms either.

When such systems need to be optimized, ad hoc methods are most commonly used, i.e. the simplest of the possible solutions, which are usually far from the optimal solutions.

For the past few years, there has been a new kind of analytics on the market called data analytics, or data scientist, a profession that has emerged in response to the need to manage the vast amounts of data that the economy is constantly producing.

Another domain is classification models that predict whether or not a given phenomenon will occur.

Machine learning models do not optimise production in any way because they are aimed at explaining phenomena or determining the outcomes of processes. Optimization methods called operational programming are designed to indicate the extreme, achievable value and explain how to obtain it.

Optimization is therefore an area reserved for a very specific type of analytics, but it's worth knowing about it because it's a picture of extreme management efficiency.

Wojciech Moszczyński — graduate of the Department of Econometrics and Statistics of Nicolaus Copernicus University in Toruń; specialist in econometrics, finance, data science, and management accounting. He specializes in the optimization of production and logistics processes. He conducts research in the area of the development and application of artificial intelligence. For years he has been engaged in the popularization of machine learning and data science in business environments.