Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2410.14316 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909354162126848 |
|---|---|
| author | Lorenz, Catherine Otto, Alena Gendreau, Michel |
| author_facet | Lorenz, Catherine Otto, Alena Gendreau, Michel |
| contents | Major players in e-commerce process dynamically incoming orders in real-time and already use advanced anticipation techniques, like AI, to predict characteristics of future orders. However, at the warehousing level, there are still no unambiguous recommendations on integrating anticipation with intelligent online optimization algorithms, nor an unbiased benchmark to assess the improvement potential of advanced anticipation over myopic techniques, as optimal online solutions are usually unavailable. In this paper, we compute and analyze Complete-Information Optimal policy Solutions (CIOSs), of an exact perfect anticipation algorithm with full knowledge of future customer orders' arrival times and items, for picking operations in picker-to-parts warehouses. We provide analytical properties and leverage CIOSs to uncover decision patterns that enhance simpler algorithms, improving both makespan (costs) and order turnover (delivery speed). Using a metric similar to the gap to optimality, we quantify the gains from optimization elements and the improvements remaining for advanced anticipation mechanisms. To compute CIOSs, we design the first exact algorithm for the Order Batching, Sequencing, and Routing Problem with Release Times (OBSRP-R), based on dynamic programming. Our analysis advises on the largely overlooked intervention - the dynamic adjustment of started batches, and the strategic relocation of an idle picker towards future picking locations. The former affected over 60% of CIOS orders and triggered consistent improvements across warehousing policies. The latter occurred before 39%-62% CIOS orders and could decrease a myopic policy's gap to optimal turnover by 4.3% on average (up to 14%). Notably, we challenge the debated concept of strategic waiting, revealing why the latter resembles an "all-in" gamble and harms both makespan and order turnover when intervention is allowed. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_14316 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | On picking operations in e-commerce warehouses: Insights from the complete-information counterpart Lorenz, Catherine Otto, Alena Gendreau, Michel Optimization and Control Major players in e-commerce process dynamically incoming orders in real-time and already use advanced anticipation techniques, like AI, to predict characteristics of future orders. However, at the warehousing level, there are still no unambiguous recommendations on integrating anticipation with intelligent online optimization algorithms, nor an unbiased benchmark to assess the improvement potential of advanced anticipation over myopic techniques, as optimal online solutions are usually unavailable. In this paper, we compute and analyze Complete-Information Optimal policy Solutions (CIOSs), of an exact perfect anticipation algorithm with full knowledge of future customer orders' arrival times and items, for picking operations in picker-to-parts warehouses. We provide analytical properties and leverage CIOSs to uncover decision patterns that enhance simpler algorithms, improving both makespan (costs) and order turnover (delivery speed). Using a metric similar to the gap to optimality, we quantify the gains from optimization elements and the improvements remaining for advanced anticipation mechanisms. To compute CIOSs, we design the first exact algorithm for the Order Batching, Sequencing, and Routing Problem with Release Times (OBSRP-R), based on dynamic programming. Our analysis advises on the largely overlooked intervention - the dynamic adjustment of started batches, and the strategic relocation of an idle picker towards future picking locations. The former affected over 60% of CIOS orders and triggered consistent improvements across warehousing policies. The latter occurred before 39%-62% CIOS orders and could decrease a myopic policy's gap to optimal turnover by 4.3% on average (up to 14%). Notably, we challenge the debated concept of strategic waiting, revealing why the latter resembles an "all-in" gamble and harms both makespan and order turnover when intervention is allowed. |
| title | On picking operations in e-commerce warehouses: Insights from the complete-information counterpart |
| topic | Optimization and Control |
| url | https://arxiv.org/abs/2410.14316 |