Watch this Ask the Expert session to learn how you can use the flexible capabilities of out-of-the-box solvers to optimize logistics routing operations and improve operational efficiency.
Watch the Webinar
You will learn:
- The different capabilities available in the SAS vehicle routing problem (VRP) solver.
- How to create a customized optimization solution by combining solvers and the flexibility of SAS Optimization procedure.
- The different network solvers available in SAS Optimization.
The questions from the Q&A segment held at the end of the webinar are listed below and the slides from the webinar are attached.
Q&A
What characteristics does the VRP network solver algorithm have in SAS Optimization?
Let's break it into two sections. In the first section let us focus on the standard Vehicle routing Problem (VRP) where a set of customer nodes are serviced from a central depot using a fleet of homogenous capacity vehicles. Note that this standard VRP doesn’t have time-windows and service time considerations. The VRP solver generates an initial primal feasible solution using a package named VRPH, and improves it using a MILP solver to find an optimal solution.
The second section of the problem has advanced rules. This is an extension of the standard VRP with time-windows for servicing customer nodes, heterogeneous vehicles, and vehicle-node eligibility rules. Currently, this problem is solved using a heuristic algorithm in our SAS VRP solver and the MILP solver is not available for improving the heuristic solution. However, the heuristic we use to solve the problem has proven to yield near optimal solutions for a variety of benchmark instances.
Again, when it comes to real world customer problems, there might be distinctive characteristics compared to the benchmark problems which we tested. The initial solution for the problem instance can be improved using additional solves using MILP, column generation algorithm etc. In todays’ talk, for the VRP problem with time-windows, we started the first phase of our solution with our VRP solver and a custom-built greedy heuristic. The second phase of the solution was a custom-built column generation algorithm to improve the solution from phase 1. The third phase was to further improve it using a compact MILP formulation.
In summary, the SAS VRP solver can handle and support solving standard VRP, VRP with time-windows, and allows us to support solution development for problems with more advanced rules.
Can I write my own heuristic/meta-heuristic algorithms in SAS Optimization?
Yes, it is more convenient to write our own custom heuristic/meta-heuristic algorithms within SAS Optimization because of the structure on which PROC OPTMODEL is built. You can define your sets, variables, and conveniently use it in do loops and for loops in your algorithm. The greedy algorithm, ACO algorithm, and the SA algorithm we discussed today were all developed within PROC OPTMODEL.
What are the benefits of writing custom optimization solutions in SAS Optimization?
Custom optimization solutions are suited for large-scale real-world problems with advanced business rules. With that said, this also depends on the type of problem being solved. For example, a large LP problem to determine distribution plans in a supply chain can be easily solved using our LP solver.
On the other hand, a large scale VRP with time-windows and advanced business rules will benefit from custom optimization solutions, by leveraging the structure of the problem. In this case, a pure MILP formulation might not yield a solution and converge as good as a custom optimization solution. Note that the SAS MILP solver is used within the iterations of the custom optimization solution. A custom optimization solution typically has an initial feasible solution (generated using a heuristic methodology) which is then further improved using a custom-built optimization algorithm.
Recommended Resources
The Network Solver - Documentation
Network Analysis and Network Optimization in SAS® Viya® - Training
What is the shortest tour that visits only once the 48 contiguous US State Capitals?
Please see additional resources in the attached slide deck.
Want more tips? Be sure to subscribe to the Ask the Expert board to receive follow up Q&A, slides and recordings from other SAS Ask the Expert webinars.
Click image to register for webinar