From Route to Resource Optimization: Augmenting Innovation with Genetic Algorithms

A genetic algorithm (GA) is an optimization algorithm built based on the natural evolutionary processes of nature. Based on the Darwinian principle of natural selection, i.e., the survival of the fittest, the concept of GAs originated from research by John Holland at the University of Michigan in 1960.^[1] Having gained popularity after the 90s, this technology now has endless potential and applications for a range of industries — research and development, gaming, healthcare, agriculture, manufacturing, logistics, finance, and much more.

When compared to conventional optimization methods (ex: gradient descent), GAs have the upper hand while processing noisy and time-consuming data. This is because they employ a derivative free optimization (when the derivative information is not used to find the optimal solutions) approach to problem-solving.^[2] Given a large data set, GAs can perform searches for the optimum set of data that satisfies given conditions and are therefore essential to data mining and pattern recognition. They have displayed high accuracy with regard to solving problems across domains and functions, including scheduling, inventory maintenance, forecasting, information security, video and image processing, and precision agriculture. Some of the industries and processes for which GAs are creating an impact are the following:

• Research and Development
• Logistics
• Manufacturing
• Finance

GA in R&D: From robotics to acoustics:

In the field of research, GAs have been applied in a variety of scenarios – from robots that can learn to walk on their own to wind turbines that can generate electricity.

GAs have been instrumental to various research projects related to acoustics as well, with applications ranging from distinguishing between sonar reflections and diverse types of objects to designing active noise control systems, which cancel sound by producing sound waves that intercept noises. Such applications can help reduce high decibel noise pollution produced during manufacturing and transport, among other use cases, for industries dealing with heavy machinery.

GAs’ success has been due to three key characteristics: they are simple to write, easy to customize, and efficient to run.^[3] Owing to their flexibility and autonomous nature, they are set to become ubiquitous in research across all disciplines, helping industries achieve cost-effective, autonomous solutions.

Intelligent route optimization: streamlining logistics.

One of the most popular applications of GAs is the traveling salesman problem.

Figure 1: Travelling Salesman Problem

This is a well-known combinatorial issue that is solved by employing unique crossover and packing procedures. GAs can be applied to assess the swiftest and most economical trips and thus effectively strategize delivery routes. As time and speed are crucial factors in the logistics industry, businesses can map the ideal delivery routes using a route planner based on GA. This can help increase efficiency for supply chain delivery as well as logistics – for example, metaheuristic architectures can optimize truck routes, which can in turn save up to $4 million per year per distribution center, on an average, and reduce carbon footprints by 28% as well. By automating such processes and making them more intelligent and efficient, GAs have revolutionized the way in which global logistics companies handle data, run operations, and cater to their customers’ needs.

Facilitating resource optimization in the manufacturing industry.

GAs can be applied in computer aided process planning (CAPP) to optimize planning for manufacturing and enable economical production at competitive rates.^[4] Currently, GAs are being used by large manufacturing companies, such as those producing aircraft, to optimize resources. For instance, the global civil aviation industry allocates around $50 billion for spare parts, which forms 75% of airlines’ inventory funds. However, only 25% of this is utilized.^[5] In such circumstances, GAs can be implemented to assess the most suitable configuration plans for spare parts. They can thus be used to streamline revenue and inventory as well as eliminate redundancies. Furthermore, GA optimization based on supply chain game decision-making models can provide quantitative assistance for signal selection and transaction decisions.^[6] Additionally, efficient packing systems can be designed with the use of GA to reduce computation times and improve high filling rates by 99%.^[7] In this way, smart machinery can be leveraged to reduce production time drastically.

GAs have also been leveraged to create thermal protective clothing for high-temperature work environments. This is done by first applying partial differential, nonlinear programming, and finite difference methods to create the overall heat transfer model. Following this, a human body burn model is obtained preliminarily and the threshold and weights of the PB neural networks are then optimized by GAs.^[8] Such attire can be worn to prevent burns, heat strokes, and other similar hazards in elevated temperature environments in manufacturing settings; GAs can therefore be leveraged to improve workplace safety and reduce the risk of injury.

Boosting profitability in finance:

GAs offer a wide range of applications in the finance sector, for instance, enabling the development of probabilistic models — which contain data in disorganized patterns — by performing reproduction, cross-over, and mutation operations on randomly generated data. The ability to deal with non-stationary data and not having to assume a specific data distribution are two of the major advantages of GAs. This method has proven to be effective and well-suited to simulating human economic behavior. GAs can also be leveraged to predict the future performance of publicly traded stocks ^[9] and other investment decisions in the real-estate market^[10] and securities trading by optimizing trading rules and creating effective business strategies.

Genetic algorithms: Future-ready applications for industries.

In the past few years, GAs have been leveraged to solve real-life complex challenges in various fields including economics, engineering, manufacturing, and R&D – identifying and eliminating redundancies, providing rapid solutions, and optimizing resources. The simple programming involved in GAs makes them easy to apply across diverse problems encountered in industries, such as planning optimal traveling routes, optimizing production processes, and predicting market trends, enabling industries to reduce carbon footprints, leverage next-gen autonomous technology, and sustain revenue in the long run.