This article, "Los Movimientos, Part II," details the application of Adaptive Large Neighborhood Search (ALNS) to solve complex pickup-and-delivery problems, a common challenge in logistics. In Part I, a smaller problem was solved using Mixed Integer Linear Programming (MILP). However, as the problem size increases, MILP's scalability becomes an issue.

ALNS is introduced as a powerful metaheuristic for larger instances. It works by iteratively destroying parts of an existing solution and then reconstructing them. The process involves selecting a destroy operator (e.g., random, worst, or related removal) to remove requests, creating "holes" in the routes. Subsequently, repair operators (e.g., greedy or regret-2 insertion) attempt to reinsert these requests while respecting all constraints, such as pickup-before-delivery, capacity limits, and time windows.

The need for ALNS arises because exact methods like MILP, while providing optimal solutions, can become computationally prohibitive for large-scale routing problems. In real-world logistics, a good, feasible solution delivered quickly is often more valuable than a mathematically proven optimum that takes too long to compute. ALNS offers a practical approach to find efficient routes within acceptable timeframes, adapting to evolving demands.