High lift movement in a storage rack system

[摘要] ENGLISH SUMMARY : With the growing competitive global economy new techniques need to be adopted for distribution of goods. Companies are searching for efficient distribution methods to improve customer service. Due to an increasing demand, distribution centres (DC) must be more versatile to keep up with a higher flow of goods. Working closely with South-Africa's largest single brand retail chain, PEP, it is determined that there is a bottle neck in the pallet storage rack system at their Durban DC. With PEP's current system the rate at which pallets are brought into the storage rack system exceeds that at which pallets can be removed from the storage rack system. This difference, in the storage and retrieval rate of pallets, means that the DC needs large temporary storage areas for pallets waiting to be stored. Furthermore, the slow rate at which pallets are retrieved from the storage rack system slows all processes down the supply chain. This study aims to nd alternative movement logics, slotting congurations and job sequences for storing and removing pallets from the storage rack system in an attempt to increase the rate at which pallets can be stored and retrieved. Five high-lift movement logics are compared namely, same side queueing without paired jobs, opposite side queueing without paired jobs, same side queueing with paired jobs, opposite side queueing with paired jobs and cyclic queueing with paired jobs. The effect of changing the sequence of jobs are also analysed for each high-lift movement logic. The job sequence is determined using heuristics that strive to nd a good sequence. Four tabu search meta-heuristics are used that respectively sequence jobs by randomly swapping jobs, systematically swapping jobs from the top, nding the worst current job pairing and improving it and lastly nding the best current job pairing and improving it. The Hungarian method is also introduced in an attempt to nd a better job sequence. Additionally, the pallet slotting problem is investigated to nd the best storage rack placement conguration based on pallet replenishment cycles. The congurations considered are column pallet slotting (fastest moving inventory (FMI) closest to the picking line), row pallet slotting (FMI closest to the ground) and a stepping conguration with the FMI on the lowest step. The movement logics, slotting congurations and job sequences are incorporated into a DC simulation program, coded in Python 3.4. Using data received from PEP's warehouse management software 10 consecutive shifts are selected for analysis. By simulating dierent combinations it is possible to nd the one that reduces the process time the most. It is found that by changing PEP's current system to one that utilises same side queueing with job pairing the total cumulative completion time is reduced by approximately 39 hours and the distance travelled by high lifts by 45.27 km. This is further improved by adding the row pallet slotting conguration and a job sequencing algorithm. Using this combination reduces the total cumulative completion time by 43 hours.

[发布机构] Stellenbosch University