Instances of the paper "The Latency Capacitated Family Traveling Salesperson Problem"
<p dir="ltr">The picking process within mega-warehouses is one of the most significant challenges in retail logistics, as it can account for more than 50% of the total cost of warehouse operations. In this study, we introduce the Latency Capacitated Family Traveling Salesperson Probl...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | , |
| Published: |
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p dir="ltr">The picking process within mega-warehouses is one of the most significant challenges in retail logistics, as it can account for more than 50% of the total cost of warehouse operations. In this study, we introduce the Latency Capacitated Family Traveling Salesperson Problem ($p$-LCFTSP) where items to be picked and the routes of the $p$ agents must be designed to minimize system latency, defined as the cumulative load carried during the picking process. One of the findings in this research is that not only the objective function is modified with respect to previous works, but the constraints have also changed. In fact, we show that the inclusion of the load latency implies, in a compact way, the sub-tour restrictions. In addition, two families of valid inequalities are also introduced to enhance the models' linear relaxation. Experimental results with the classical benchmarks show the importance of the latency incorporation.</p><p><br></p><p dir="ltr">Instance summary: To test the performance of our models, we use the 21 FTSP instances from [1] based on the instances of the classical TSPLIB benchmark with up to 1001 nodes (https://softlib.rice.edu/tsplib.html). To generate the $p$-LCFTSP instances, we maintain the number of nodes but, depending on their size, set 3, 4, 5, or 10 different families. The demand also depends on the size of the node set and varies between 6 and 60 items. The weights are uniformly taken from [1,5]. The number of agents also depends on the node set size and goes from 1 to 11. The distance matrices are those provided by TSPLIB. Note that for the $p$-LCFTSP, the set of instances was designed to have some instances with agents of constant capacity, and the weight for all families is 1. It also has instances where the number of agents is high, but their capacities are relatively small, and vice versa. Finally, there are instances with heterogeneous capacities and different family weights. </p><p dir="ltr">[1] Morán‐Mirabal, L. F., González‐Velarde, J. L., & Resende, M. G. (2014). Randomized heuristics for the family traveling salesperson problem. <i>International Transactions in Operational Research</i>, <i>21</i>(1), 41-57.</p><table><tr> </tr><tr><td></td></tr></table><p><br></p> |
|---|