Banu Yetkin Ekren

Simulation based experimental design to identify factors affecting performance of AVS/RS

We perform a simulation based experimental design for automated unit-load (UL) storage and retrieval systems based on autonomous vehicle technology to identify factors affecting their performance. First, we select the best combination of numbers of lifts and vehicles from pre-defined scenarios that are the key components of the system. Then, we apply design of experiment (DOE) for a system with this combination of lifts and vehicles and for various arrival rates. The factors considered in the DOE include: dwell point policy, scheduling rule, input/output (I/O) locations and interleaving rule. Three different responses, the average cycle time for storage and retrieval transactions, average vehicle utilization, and average lift utilization, are considered. However, because the ANOVA assumptions are not met for the average cycle time response, an inverse transformation method is applied on this response. The results show that there is three-way interaction effect on each response at a 95% confidence level. After determining the main and the interaction effects, a Tukeys test analysis is completed on the responses. We utilize data from a warehouse in France that utilizes the autonomous vehicle storage and retrieval system.

Simulation-based regression analysis for the rack configuration of an autonomous vehicle storage and retrieval system

In this paper, a simulation-based regression analysis for the rack configuration of an autonomous vehicle storage and retrieval system (AVS/RS) is presented. The aim of this study is to develop mathematical functions for the rack configuration of an AVS/RS that reflects the relationship between the outputs (responses) and the input variables (factors) of the system under various scenarios. In the regression model, we consider five outputs: the average cycle time of storage and retrieval transactions, the average waiting time for vehicle transactions, the average waiting time of vehicles (transactions) for the lift, the average utilisation of vehicles and the average utilisation of the lifts. The input variables are the number of tiers, aisles and bays that determine the size of the warehouse. Thirty regression models are developed for six warehouse scenarios. The simulation model of the system is developed using ARENA 12.0 commercial software and the statistical analyses are completed using MINITAB statistical software. Two different approaches are used to fit the regression functions–stepwise regression and the best subsets. After obtaining the regression functions, we optimise them using the LINGO software. We apply the approach to a company that uses AVS/RS in France.

An Approximate Solution for Semi-Open Queueing Network Model of an Autonomous Vehicle Storage and Retrieval System

We present an analytical model for an autonomous vehicle storage and retrieval system (AVS/RS). The system is modeled as a semi-open queueing network (SOQN). An SOQN consists of customers, a secondary resource and servers. Each arriving customer is paired with the secondary resource. The two visit the set of servers required by the customer in the specified sequence. In the context of an AVS/RS, storage/retrieval (S/R) transactions are customers and the autonomous vehicles are the secondary resources. If an S/R transaction requires a vertical movement, it uses a lift. The lifts and horizontal travel times to and from a storage space are modeled as servers. First, we define all possible scenarios for storage and retrieval transactions and their occurrence probabilities. Second, we derive general travel times of vehicles and lifts by considering all possible locations of the two devices based on the predefined storage and retrieval scenarios, defined in step 1. Third, each scenario is modeled as a customer type and these customer classes are aggregated into a single class. Thus, we model the system as a single-class, multiple-server, SOQN. Finally , we solve the SOQN using an approximate method and obtain the performance measures. We apply the method to analyze a warehouse in France that utilizes AVS/RS.

Matrix-geometric solution for semi-open queuing network model of autonomous vehicle storage and retrieval system

In this paper, we model the autonomous vehicle storage and retrieval system (AVS/RS) as a semi-open queuing network (SOQN) and apply a matrix-geometric method (MGM) for analyzing it. An AVS/RS is an automated material handling system for the high-rise pallet storage area of a warehouse and allows pallets to be stored and retrieved quickly and efficiently from their storage locations. It is an alternative to the traditional crane-based AS/RS (automated storage and retrieval system). A combination of lifts and autonomous vehicles store pallets into and retrieve them out of their respective rack storage locations. The crane based AS/RS typically utilizes aisle-captive, mast-mounted cranes that can access any storage location in an aisle via horizontal movement of the mast and vertical movement of the crane on the mast. In an SOQN, it is assumed that an arriving job or customer is paired with another device and the two visit all the stations that must process the job in the appropriate sequence. After all operations are completed on the job, it exits the system, but the device returns back to a device pool and awaits the next customer. Sometimes a job may have to wait for a device to arrive at the pool or a device may have to wait for a job to arrive. Although closed queuing networks (CQNs) and open queuing networks (OQNs) model systems that require pairing of an incoming job with a device, unlike the SOQN, they ignore the time that a device waits for a job or the time that a job waits for a device. In the context of an AVS/RS, the jobs correspond to storage/retrieval (S/R) transaction requests and the autonomous vehicles (AVs) correspond to the devices. Because an AV may sometimes have to wait for an S/R transaction or vice versa, we model the AVS/RS as an SOQN. We build the queuing network by deriving general travel times of pre-defined servers. We model the AVS/RS system as a single-class, multiple-server, SOQN. Then, we solve the network using the MGM and obtain its key performance measures. We apply the MGM technique for solving the SOQN model to a warehouse in France that uses AVS/RS.

A simulation based experimental design to analyze factors affecting production flow time

Abstract In this paper we analyze and evaluate the effects of some pre-defined process parameters on the performance of a manufacturing system. These parameters include two different plant layout types, namely functional layout (FL) and cellular manufacturing layout (CL), as well as scheduling rule, machine downtimes, batch sizes, and transporter (interstage transporters) capacities. First we employ simulation to evaluate the effects of these factors on the performance of the system and then conduct designed experiments to set the best levels for these factors. The performance evaluation function is defined in terms of the average flow time of all the part types through the manufacturing system. Arena 10.0 simulation software and SPSS 9.0 statistical package are used to measure the main effects and interactions between these factors. This work demonstrates that various manufacturing parameters should be considered jointly when designing or redesigning a facility because setting different levels for parameters can considerably affect the performance of a facility.

Simulation based regression analysis for rack configuration of autonomous vehicle storage and retrieval system

In this study, a simulation based regression analysis for rack configuration of an autonomous vehicle storage and retrieval system (AVS/RS) is presented. We develop a mathematical function for rack configuration of an AVS/RS that reflects the relationship between the output (response) and the input variables (factors) of the system. In the regression model, the output is the average cycle time for storage and retrieval and the input variables are the number of tiers, aisles and bays that determine the size of the warehouse. The simulation model of the system is developed using ARENA 12.0, a commercial software. We use MINITAB statistical software to complete the statistical analysis and to fit a regression function. Two different approaches are used for developing the regression analysis — stepwise regression and the best subsets. We optimize the regression function using the LINGO software. We apply this approach to a company that uses AVS/RS in France.

Simulation based optimization of multi-location transshipment problem with capacitated transportation

In this study, a single-item two-echelon inventory system where the items can be stored in each of N stocking locations is optimized using simulation. The aim of this study is to minimize the total inventory, backorder, and transshipments costs, based on the replenishment and transshipment quantities. In this study, transshipments which are the transfer of products among locations at the same echelon level and transportation capacities which are the transshipment quantities between stocking locations, are also considered. Here, the transportation capacities among the stocking locations are bounded due to transportation media or the locations¿ transshipment policy. Assuming stochastic demand, the system is modeled based on different cases of transshipment capacities and costs. To find out the optimum levels of the transshipment quantities among stocking locations and the replenishment quantities, the simulation model of the problem is developed using ARENA 10.0 and then optimized using the OptQuest tool in this software.

Graph-based solution for performance evaluation of shuttle-based storage and retrieval system

The aim of this study is to provide a graph-based solution for performance evaluation of a new autonomous vehicle-based storage and retrieval system, shuttle-based storage and retrieval system (SBS/RS), under various design concepts. By the graph-based solution, it is aimed the decision-maker (i.e. warehouse manager) evaluates a pre-defined system’s performance promptly and decides on the correct design concept based on his/her requirements from thousands of alternative design scenarios of SBS/RS. The design concepts include number of bays (NoB), aisles (NoA) and tiers (NoT) for the rack design and arrival rate of storage/retrieval (S/R) transactions to an aisle of the warehouse (AR). The performance of the system is evaluated in terms of average utilisation of lifts and average cycle time of S/R transactions. Simulation is utilised for the modelling purpose. Seven NoT, seven NoB and six AR scenarios are considered in the experiments. Hence, 294 experiments are completed to obtain the graphs. By this study, to the best of our knowledge it is the first time a graph-based solution including comprehensive design concepts of SBS/RS is presented.

Method for evaluating the throughput performance of shuttle based storage and retrieval systems

In this paper a method for throughput performance calculation of shuttle based storage and retrieval systems (SBS/RS) is presented. SBS/RS represent a new technology in automated storage and retrieval systems. Since it is important to design SBS/RS right the first time due to the relative inflexibility of the physical layout, we provide a proposed method for the throughput performance calculation of these systems. The performance of the system is considered as a throughput capacity of the SBS/RS as a whole.

Approximate analysis of load-dependent generally distributed queuing networks with low service time variability

In this paper, we present an approximate method for solution of load-dependent, closed queuing networks having general service time distributions with low variability. The proposed technique is an extension of Maries (1980) method. In the methodology, conditional throughputs are obtained by an iterative procedure. The iterations are repeated until an invalid result is detected or no improvements are found. We demonstrate the performance of the technique with 10 different examples. On average, the solutions have 5% or lower deviations when compared to simulation results.