Exploring the Application of Swarm Intelligence Algorithms to Nonlinear and Non-Convex Problems in the Real-World

In the real world, when solving problems based on advanced artificial intelligence (AI) technology, the problem is usually first transformed into an optimization model and corresponding algorithms are designed to solve it. Usually, such problems belong to non-convex nonlinear problems with numerous parameters, which are usually solved using swarm intelligence algorithms. Therefore, in this study, we summarized the three most common non convex nonlinear problems in the real world, including drone path planning, desert crossing problem, and workshop scheduling problem. We have summarized the three problems encountered and established corresponding mathematical models. The problem of drone path planning is based on highway or infrastructure detection tasks. Crossing the Desert is a real-life game. The workshop scheduling problem is based on the background of automobile factories. On this basis, we developed an improved dragonfly algorithm (DA) and applied the improved DA algorithm to three mathematical models.