The single vehicle pickup and delivery problem with time windows is an important practical problem, yet only a few researchers have tackled it. In this research, we compare three different approaches to the problem: a genetic algorithm, a simulated annealing approach, and a hill climbing algorithm. In all cases, we adopt a solution representation that depends on a duplicate code for both the pickup request and its delivery. We also present an intelligent neighborhood move, that is guided by the time window, aiming to overcome the difficult problem constraints efficiently. Results presented herein improve upon those that have been previously published.Keywords Pickup and delivery problem with time windows · Dial-a-ride · Vehicle routing · Genetic algorithms · Simulated annealing · Hill climbingThe pickup and delivery problem with time windows (PDPTW) is an important practical problem that is likely to assume even greater prominence in the future. Current concerns over global warming, resource depletion and the social impact of traffic congestion and pollution (and resulting legislation, increase in cost, and changes in public perceptions) are driving companies, government organizations and researchers to improve the efficiency of logistics and distribution operations. In addition, the rapid growth in parcel transportation as a result of e-commerce is likely to have an increasing impact. More cooperation between manufacturers, shippers and carriers in supply chains could greatly reduce the environmental impact of transport (the number of M.I. Hosny ( ) ·