Determination of picking policies in robotic compact bin picking storage systems with artificial bee colony algorithm

In this thesis, the subjects of warehouse design and operations management are investigated by considering technological developments. Robotic Compact Bin Storage (RCBS) systems, which are among the systems based entirely on automation, have been examined in detail, and notations representing the structural features of the system have been produced. Considering the differences between the systems, pickers to parts and parts to pickers have begun to separate today. The order-picking process of the RCBS system has been analyzed in detail, and results that will be beneficial to both researchers and practitioners have been revealed. In this thesis, new methods and concepts are presented to model order picking systems and analyze their performance in the RCBS system. Specifically, in this thesis, SLAP is dynamically investigated in multiple periods with the help of heuristic and metaheuristic algorithms. In addition, significant savings have been achieved in total travel times and labor hours in terms of OPP in RCBS systems. Significant gains in collection times have been achieved by testing the rearrangement approach at different levels. At the same time, the structure of the RCBS system, which provides higher collection efficiency compared to the traditional storage methods used, has been examined in detail. The effect of the improvements on the system has been tested using the simulation technique. The concepts and models discussed in this thesis offer warehouse researchers and practitioners new perspectives to understand and improve order picking system performance. Picker to parts systems where the pickers go to the address of the relevant product and transport the products to the loading area with the help of various equipment in corridors with a certain plan. Forklifts, pallet trucks or special collection baskets are used as collection equipment in these warehouses. The systems in which the picker moves to the product are also called traditional manual warehouses in the literature. Studies on traditional manual warehouses are frequently encountered in the literature. In parts to picker systems, an automatic vehicle takes the product from its address and delivers it to the order picker. After the products are received by the order picker, the remaining products, if any, are taken back to the address. The picking elements used in these systems are faster than picker to parts systems. For this reason, the order completion time is shortened in parts to picker systems With the development of parts to picker systems, wide corridors are no longer needed to use traditional picking equipment. In this way, more products can be stored in warehouses. These new-generation warehouses are called compact warehouses. New-generation compact warehouses are advanced in information systems, where collecting equipment such as robots, shuttles, and lifts are used. RCBS systems, an example of new-generation compact warehouses, increase storage space efficiency. In these systems, shipment may continue during operations such as installation and expansion. Human-based product losses are eliminated. Measures can be easily arranged according to demand change or seasonal effects. They are more green-sensitive systems thanks to their very low energy use. The Storage Location Assignment Problem (SLAP), one of the several sorts of problems that can be examined in the warehouse, will be the focus of this thesis. In order to improve the effectiveness of the order picking system, the performance of the rearrangement strategy is investigated. Dynamic SLAP is the problem of determining the locations of each product coming to the warehouse by the general efficiency understanding and according to the strategic targets. The traditional methods used to solve this problem consist of a simple set of rules aimed at storing incoming products and assigning them to their location. Instead of the traditional methods used in warehouses, the Artificial bee colony algorithm, which is among the artificial intelligence techniques, was used to solve SLAP within the scope of this study, and a new warehouse location assignment heuristic was produced. A 0-1 integer linear mathematical model has been established, which minimizes the time spent during the picking and loading operations. Due to the NP-hard nature of the mathematical model, the Artificial Bee Colony (ABC) algorithm, one of the meta-heuristic algorithms, produces near-optimal solutions in a short time by imitating the swarm intelligence of honey bee hives, has been used. In the ABC algorithm, the food sources show possible solutions, and the information about the quality of the food source is calculated with the membership function. An order route is the sum of the operations performed between different points during the warehouse order-picking process. Order times can be reduced by placing products found in the same order in close places in the warehouse. For these purposes, the Quadratic form of the Q-SLAP mathematical model was produced. Congestion in the picking area increases order delivery time and order completion time. Finding solutions to reduce congestion and putting them into practice provides scientific and managerial contributions. Congestion in the picking area increases order lead time and completion time. Finding solutions to reduce congestion and putting them into practice provides great advantages for OPP. Re-arrangement approach has been developed in the literature, which will provide the opportunity to be operationally prepared by noticing such situations in advance. In the years when this approach was tested, due to the use of traditional warehouses, the effects of the measures taken on labor costs and order delay were at an acceptable level. The re-arrangement approach, which can only be performed with robots in RCBS systems, has been analyzed in terms of various performance criteria. The components that make up the cost of order picking in the warehouses are examined. It has been revealed that the labor cost used in the picking process is one of the highest costs of the warehouse. In traditional warehouses, rearrangements require more workers or overtime. However, in the new generation compact warehouses, the picking process is done by robots. For this reason, the rearrangement process can be unmanned, outside working hours, or simultaneously with the order-picking process. In the RCBS system, robots can visit more than one point for an order and queue in front of the stations to prepare different orders. Exact analysis of such systems is difficult today. A simulation model is developed to dynamically predict the performance of the RCBS system (for example, the average completion time of the order in the system, the average total picking time, the average total reorder time, the average total loading time, the simulation time, etc.) under different storage strategies. Within the scope of this study, it has been shown that the rearrangement process in the RCBS system reduces the order-picking time using the simulation technique.