Example of an order batching algorithm for picking

Example of an order batching algorithm for picking

Order batching is one of the methods used during the picking process. The method involves combining a certain number of picking orders into batches and simultaneously picking all orders within a given batch. The purpose of this action is to increase the efficiency of the picking process by:
– shortening the total length of picking routes that must be covered to collect all items located in the analyzed group of orders – as a result, picking can be carried out faster or with fewer employees and internal transport resources,
– fuller utilization of the capabilities of the internal transport resources used (especially when small items are being picked and there are relatively few items on the picking orders).

In general, in such cases, the internal transport resource used has the capacity to simultaneously carry items from more than one base picking order. In general, order batching algorithms can be divided into three main subgroups:

1. Priority rule based algorithms. This is a group of the simplest algorithms; the two most commonly used variants are:
– “first come, first served” rule – orders are added to the batch one by one (in chronological order of appearance/arrival) until the next order exceeds the transport capacity of the adopted transport resource. In such a case, the current batch is closed, and a new batch is opened with that specific order,
– parallel batch rule – several (a dozen, several dozen) parallel batches are created, and the next order is added to the batch with the lowest number that simultaneously still has sufficient space (first fit rule) or to the batch that has the space closest to the size of the given order (best fit rule).
Such algorithms are relatively simple both in terms of their implementation and the computational time required for their execution. No additional calculations are required here, and the analysis can be performed solely on the basis of picking orders.

2. Savings algorithms – the operation of such algorithms consists of comparing how much distance will be saved by creating specific batches of orders and searching for the most favorable combination of orders included in these batches (and the most favorable number of batches). The operating principle of savings algorithms is shown in Figure 1, where A, B, and C are picking routes for individual orders, while D is the picking route for combined orders A, B, and C. The combined route (D) is longer than (or at most equal to) each of the routes determined for orders A, B, and C, while at the same time, the length of route D is significantly smaller than the sum of the lengths of routes A, B, and C. For the above case, the “saving” is the difference between the length of route D and the sum of the lengths of routes A, B, and C. The goal of savings algorithms is to maximize the sum of such savings. For proper operation, such an algorithm requires determining picking routes (and their lengths) for each variant of order batching. The number of possible variants is also very high – for example, for 100 picking orders, almost 4 million different 4-element order batches can be created. Even using heuristic picking route determination algorithms, the calculations necessary for all possible solutions will be very time-consuming, and due to computational complexity, they will become practically unfeasible when using an optimal picking route determination algorithm. Additionally, apart from the number of required calculations, order batching lengthens the picking list, which can have a significant negative impact on the performance of such an algorithm. For this reason, the optimal solution to the order batching problem for savings algorithms is considered an NP-hard problem.

3. Seed algorithms (based on selecting a “seed” order and “additional orders” according to certain criteria) – this group of algorithms acts as an intermediate link in terms of efficiency and calculation speed between the two types of algorithms mentioned above. The general operating principle of this group of algorithms can be described as follows:
a) select an initial order (according to the adopted criterion), creating a new batch of orders,
b) select an additional order (according to the adopted criterion) that can be included in the batch created in point 1,
c) repeat point 2 as long as there are unbatched orders that can be included in the given batch or the given batch has reached its maximum capacity,
d) repeat points 1-3 until all orders have been batched or no more batches can be created (according to the adopted criterion/criteria).

 

The full article is available below in PDF format.

Source: Logistyka Magazine

L2017-2s34.pdf