Track: High School STEM Competition

Abstract

Food and parcel delivery routing is a major problem area due to its dynamism and complexity, exacerbated in recent years by growth in demand. The problem was modelled as a variation of the OP, a classical combinatorial optimisation problem, specifically the OTOP, involving a team of vehicles where each vehicle has a time budget to serve locations. A hybridisation of the metaheuristics Simulated Annealing and Iterated Local Search (SAILS) was constructed, able to escape local optima by sometimes accepting worse solutions. A website was also created to showcase the algorithm and computational tests were conducted on instances with varying parameters to simulate the large breadth of real-world scenarios. SAILS had higher profit than ILS and the initial solution and was able to acquire 58.8% of the profit increase over 20 minutes of runtime within 5 seconds, showing efficiency. However, the low improvement of SAILS over ILS shows SAILS has much room for improvement. Possible future work includes creating a more representative model with fewer assumptions and improving the algorithm through methods like reheating, genetic algorithms, and more precise parameter tuning. In short, a SAILS pathfinding algorithm was constructed to effectively and efficiently tackle delivery routing.