A Linear Genetic Programming Approach for the Internet Shopping Optimization Problem with Multiple Item Units (ISHOP-U)

Abstract

Evolutionary computation (EC) is a broad field of artificial intelligence where evolutionary processes inspire algorithms, such as artificial immune systems, inspired by the evolution of acquired immune systems. The predominant approach in EC is Evolutionary Algorithms (EAs), inspired by the evolution of Darwin’s natural species. A different approach is Evolutionary Programming (EP), which, instead of evolving individuals representing the problem decision variables (chromosomes), evolves programs, which code instructions, and executing those instructions generates a solution. Genetic Programming (GP) is an approach analogous to Genetic Algorithms (GAs), but it differs in that it works over programming instructions instead of decision variables. Although GP is an exciting approach, it is more complicated to implement due to the necessity of managing tree data structures. Linear Genetic Programming (LGP) is more straightforward than traditional GP, without the need for tree data structures. This chapter shows a proof of concept to implement LGP to evolve programs for the Internet Shopping Optimization Problem with multiple item Units (ISHOP-U), an NP-Hard optimization problem. Readers can easily implement the proposed approach and produce Linear Genetic Programming algorithms for other problems.