Shashank Jain - Nature-Inspired Optimization Algorithms with Java: A Look at Optimization Techniques

True optimization is the revolutionary contribution of modern research to the decision process.

George Dantzig, American scientist

This chapter introduces the optimization techniques, focusing on those that are metaheuristic/nature-inspired. You learn how to benchmark these techniques and the types of benchmark functions that are available.

Every day in business, government, and our personal lives, we decide how to best use available resources, including time and money. It can be difficult to decide which items to buy or which routes to take to work.

For organizations dealing with supply chains, this challenge is worse. For example, they must address how to load packages on a fleet of trucks, which routes drivers should take to make deliveries in the least amount time, or how to assign crews and aircraft to airline flights as they move across the country throughout the day.

Decisions on allocating limited resources to different uses when there are many options and interrelationships are prime candidates for optimization . In optimization, you need to quantitatively define a model that specifies all the ways, times, or places your resources may be allocated and all the significant constraints on resources and uses that must be met. Then determine the solution that meets the best allocation of resources.

Quantified decision variables are the number of resources allocated to each use; for example, the number of employees working each shift or the number of packages loaded onto each truck. To determine what best means, you must define a quantity called the objective that can calculate the value of the decision variablessuch as minimizing costs, maximizing profits, or minimizing production time.

Since any resource is subject to some limitations, to complete the model, you must define each constraint or limit the ways resources may be allocated that reflect real-world situations. Usually, these are simple constraint limits calculated from the decision variables, such as a maximum of 150 hours to complete the tasks.

For example, a shopkeeper wants to keep items that maximize profit or to lower inventory costs. In this case, the decision variables are the items to keep. A constraint might be the amount of capital. Profit is the objective function.

Another example is a stock investment portfolio. The objective function is to maximize wealth. The decision variables are the individual stocks. A constraint is the amount of money the investor has.