Browsing by Author "Tu, Yiliu"
Now showing 1 - 9 of 9
Results Per Page
- ItemOpen AccessA Grasshopper Optimization-Based Approach for Task Assignment in Cloud Logistics(2020-04-04) Xu, Lan; Tu, Yiliu; Zhang, YutingA framework for the algorithm-based CL platform is established, based on which, the operational mode of it is described in detail. An integrated logistics task assignment model is built to optimally match logistics service resources and task of large scale in the algorithm-based CL. Particularly, an improved grasshopper optimization-based bitarget optimization algorithm (GROBO) is proposed to solve the biobjective programming model for service matching in CL. The case of Linyi small commodity logistics is taken as an application. Simulation results show that the proposed GROBO provides better solutions regarding to searching efficiency and stability in solving the model.
- ItemOpen AccessDiscrete Choice-based Equilibrium Modeling of Supply Chain Network with Conflicting Objectives and Demand Uncertainty(2021-01-05) Ma, Jun; Tu, Yiliu; Nault, Barrie R.; Patterson, Raymond A.; Li, Simon; Ma, YongshengThis PhD thesis discusses several equilibrium problems in supply chain network integration and specifically concentrates on the importance of collaboration under conflicting objective decision-making and uncertainty management in a supply chain network. In particular, discrete choice models are introduced into supply chain equilibrium models to bridge the conflicting objective decision-making in supply chains and customer preference. Furthermore, numerical examples are provided for model illustration, managerial insights, and algorithm performance. Finally, it attempts to explore the tradeoffs between the operation costs, service level, and time issues in a supply chain, considering customer preference and demand uncertainty. Two conventional assumptions used in both the supply chain network equilibrium model and the newsvendor model are generalized and relaxed in order to obtain more general solutions and methods. First, this PhD thesis adopts Sheffi’s equilibrium condition assumption rather than Wardrop’s network equilibrium condition assumption. Discrete choice models are introduced into supply chain equilibrium models. A probabilistic fashion is used to describe customer choice behavior, because all factors affecting customer choice behavior cannot be observed completely. It assumes equilibrium will be reached when no customer believes that his utility can be improved by unilaterally changing products (or services) provided by supply chains. This assumption is not easy to use in practice, except it can be characterized and formulated as equilibrium conditions mathematically. Next, the equilibrium conditions are formulated as multinomial logit- and newsvendor model-based variational inequalities. Second, the assumption in the newsvendor model that unmet demand is lost implies that customers are stockout neutral. This study assumes the customers are stockout aversion. This work generalizes the implied assumption in the newsvendor model and introduces it into the supply chain equilibrium model. This PhD thesis has several contributions to the supply chain network integration with a focus on the collaboration under conflicting objective decision-making and uncertainty management in a supply chain network. First, discrete choice models are incorporated into supply chain equilibrium models to bridge the conflicting objective decision-making in supply chains and customer preference. It brings several technical problems. A variety of corollaries, theorems, and propositions are provided to illustrate the models and problems. Hence, the model can optimize supply chain profits without multi-objective conversion at the firm level. Second, heterogeneous customers’ discrete choice behaviors are considered in the model. Most existing supply chain network equilibrium models are capable of providing equilibrium solutions in a supply chain network, only under the assumption that customers are homogeneous. This assumption can be extended to heterogeneous customers by using a supply chain network economic or equilibrium models integrated with the multinomial logit model and data at the level of the customer individual. Third, this work assumes that customers are stockout aversion and introduce the newsvendor model to solve the issue that uncertain demand depends on both price and service level in a supply chain network for joint products. The newsvendor model-based variational inequality problems are given to formulate the equilibrium conditions.
- ItemOpen AccessDynamic Price Quotation in a Responsive Supply Chain for One-of-a-Kind Production(Elsevier, 2012-09) Nault, Barrie R; Zhang, Jian (Ray); Yonag, Wensheng; Tu, YiliuThis paper studies the setting in which a one-of-a-kind production (OKP) firm offers two types of orders (due-date guaranteed and due-date unguaranteed) at different prices to the sequentially arriving customers, who are also OKP production firms. The prices for two types of orders are quoted to each customer on its arrival. We study two problems in this setting. First, we model a dynamic pricing strategy (DPS) and compare our DPS with a constant pricing strategy (CPS). Through a numerical test, we show that both the firm and its customers are better off when our DPS is employed, so that the DPS improves overall performance of the supply chain. Through an industry case, a custom window production firm, we show how to apply the proposed DPS when products are complex. We also develop a method to adaptively estimate the firm's available capacity, the number of future arrivals and the distributions of the customers' willingness to pay and impatience factor. The simulation result shows that, when multiple distribution parameters are unknown, the proposed parameter estimating method results in estimates close to the true values.
- ItemOpen AccessDynamic Pricing and Scheduling for the Coordination of a One-of-a-kind Production Supply Chain(2013-09-13) Zhang, Jian; Tu, YiliuIn this thesis work, we study the dynamic pricing strategy (DPS) for different cases and its influences on supply chain coordination. First, we study a DPS for a one-of-a-kind production (OKP) firm with two classes of orders (due-date guaranteed and due-date unguaranteed). We model the DPS using Bellman equation and compare it with a static pricing strategy (SPS). Second, we study the pricing problem for a third-party-logistics (3PL) provider that provides warehousing and less-than-truckload (LTL) transportation services. We develop a stochastic-nonlinear-programming (SNLP) model which computes the optimal freight rates for different delivery dates incorporating the 3PL provider's current holding cost and available transportation capacity. We develop an adjusted multinomial logit (MNL) function to predict customer choices so that our SNLP model can obtain near optimal freight rate settings. Finally, we study dynamic pricing based on a practical OKP firm which is currently employing a SPS. For the three cases, we show the increase of the price-setting firm's profit, customer and social welfare when DPS is employed through simulation, and consequently show the DPS's influence on the performance of the supply chain. We also develop a scheduling method for a manufacturer whose suppliers offer different delivery times at different prices. We abstract the problem to a one-machine scheduling problem which is featured by: (a) the release date of each job is compressible and stochastic, (b) each job has to be delivered before its due date (deadline) and (c) the manufacturer can expedite the production with costly overtime. The target is to minimize the total cost including the compressing cost and the overtime production cost. We coin a concept of a job's late-release-impact factor (LRIF) and we propose a LRIF based heuristic algorithm. Through the numerical test, we shows that the LRIF based algorithm can obtain a better schedule comparing to the ones that are commonly used in practice. By this thesis work, we are trying to integrate an OKP supply chain, which is critical to reduce the cost of OKP companies.
- ItemOpen AccessAn efficient heuristic for adaptive production scheduling and control in one-of-a-kind production(Elsevier, 2011) Nault, Barrie R; Li, Wei; Xue, Deyi; Tu, YiliuEven though research in flow shop production scheduling has been carried out for many decades,there is still a gap between research and application especially in manufacturing paradigms such as one-of-a-kind production(OKP) that intensely challenges realtime adaptive production scheduling and control.Indeed,many of the most popular heuristics continue to use Johnson’s algorithm(1954) as their core. This paper presents a state space(SS)heuristic,integrated with a closed-loop feedback control structure,to achieve adaptive production scheduling and control in OKP. Our (SS) heuristic,because of its simplicity and computational efficiency,has the potential to become a core heuristic.Through a series of case studies,including an industrial implementation in OKP,our SS-based production scheduling and control system demonstrates significant potential to improve production efficiency.
- ItemOpen AccessImprovement of Turret Punch Tool Layout for the Production of Nested Parts(2017) Anderson, David C; Tu, Yiliu; Balakrishnan, Jaydeep; Freiheit, Theo; Xue, DeyiSheet metal parts processed by CNC turret punches are often grouped together onto single sheets of material, known as nests, in random combinations based on current demand. The content and configuration of each nest is highly variable, resulting in unique hole locations and quantities. A hybrid genetic algorithm (HGA) is presented for the development of a robust turret layout given a set of parts with known tool requirements and flexible operation sequences. HGA population members are improved through an iterative local search heuristic that alternately considers part operation sequences and turret layout. Improved members replace their unimproved predecessors in the population. HGA solutions are tested for robustness using a modified form of the Layout Configuration Robustness Index (LCRI). The HGA solutions are shown to offer a statistically significant decrease in total turret rotation distance when compared to a population of randomly generated layouts.
- ItemOpen AccessIntegrated Project and Supply Chain Management in Well Drilling Process(2014-11-14) Chen, Xiuyi; Tu, YiliuThis thesis work provides a mixed integer programming model to help integrating the drilling operation and supplier selection in well drilling process of oil/gas production. The appropriate decisions on the services orders are taken based on three criteria including service duration, cost and timely deliverance. The schedule of drilling operation is based on regular working time and overtime. The research outcomes provide the optimal or rational solutions for the decisions on: supplier selection, regular working time vs. overtime planning for each activity, and the total project duration with the minimum total project cost. The two typical drilling operation project cases from a local oil/gas company are collected and conducted to validate the feasibility and effectiveness of the model. In the thesis, the conflicts and trade-offs on the business profits and project time control between the operator company and its suppliers are also discussed. To solve the problem resulted from divergent positions between the operator company and its suppliers, the sharing risk and incentive contract are suggested to be adopted by the operator companies in oil/gas production from other manufacturing research and applications. In short, this study is novel and beneficial for drilling project management as it could improve the performance of drilling operations and to integrate the activities of the suppliers.
- ItemOpen AccessModelling and Analysis of the Axial Vibration on BHA of Horizontal Drilling Rig(2018-04-30) Xu, Jingxuan; Tu, Yiliu; Xue, Deyi; Li, Leping; Fapojuwo, Abraham O.Horizontal drilling technology is a widely applied well drilling technology due to its capability of reducing drilling time and cost, and at same time increasing the production. In practice, oil and gas production companies employ rotary steerable system or positive displacement motor to steer the drill bit to follow a pre-planned well trajectory along a desired direction. Therefore, the efficiency and accuracy are the main concerns for the improvement of horizontal drilling technology and drill string vibration is the factor to influence these two major concerns. This thesis illustrates a dynamic mathematical model for the vibration of the bottom hole assembly (BHA) which is an important part of drill string. The mathematical model considers the friction between the wellbore and the BHA and the effect of drilling fluid. It investigates the major influencing factors on the vibration, such as weight on bit, friction coefficient, viscous damping coefficient, and the number of stabilizers. The dynamic mathematical model is validated by the finite element simulation and analysis. Moreover, it can monitor the deformation of BHA real time and help drillers adjusting the parameters when the well path deviates from the original planned trajectory.
- ItemOpen AccessOptimal Design of Reconfigurable Products Considering Product Configurations and Reconfiguration Processes(2018-08-23) Gadalla, Moustafa; Xue, Deyi; Ramirez-Serrano, Alejandro; Tu, Yiliu; El-Sheimy, Naser; Ma, YongshengA reconfigurable product is used as several products for delivering different functions through reconfiguration processes to change between product configurations during its utilization stage. Despite the research efforts devoted to design and manufacture of reconfigurable products, influences of reconfiguration processes have not been considered in the design of reconfigurable products. Since different designs can lead to different reconfiguration processes, and both the product configurations and reconfiguration processes influence the evaluation measures of the product, an optimal reconfigurable product design method is developed in this research considering both product configurations and reconfiguration processes.First, a generic design AND-OR tree is used to model design candidates and design parameters based on the same requirements, and generic process AND-OR graphs are employed to model reconfiguration process candidates and process parameters. Specific design and reconfiguration process candidates are created from the generic design AND-OR tree and the generic process AND-OR graphs through tree/graph based search. Based on evaluations to different design candidates and reconfiguration process candidates, the optimal design solution is identified using a multi-level and multi-objective optimization method. The optimization is conducted at two different levels: the candidate optimization to achieve the optimal design/process candidates, and parameter optimization to identify the optimal design/process parameter values. Second, the efficiency of the developed optimization method is improved when the numbers of design/process candidates are large. A heuristic is developed to evaluate nodes of the design/process tree/graphs considering the importance of these nodes to the optimal design, and some nodes that are unlikely to achieve the optimal solution are pruned from the tree/graphs. To further improve the optimization efficiency, a heuristic function is developed to rank the created design solutions to identify the top-ranked solutions for detailed optimization. Third, a new extended analytic network process (EANP) method is developed to determine weights of the design/process nodes for improving the quality of the developed optimization method considering dependency relations among descriptions of design/process candidates and evaluation criteria. In addition, by selecting the evaluation criteria in the EANP method with the similar evaluation measures in the optimization process, better top-ranked solutions can be identified for detailed optimization.