Customer Classes Research Articles

Assortment Optimization and Pricing Under a Nonparametric Tree Choice Model

We consider assortment and pricing problems when customers purchase products according to a nonparametric choice model. Each customer arrives with a preference list and will purchase the highest-ranking offered product in her preference list. We assume the set of customer classes is derived from paths in a tree, in which the order of nodes visited along each path gives the corresponding preference list. First, we study assortment problems, in which the goal is to find which products to offer to maximize expected revenue. We give a dynamic programming solution, which can be extended to versions of the assortment problem in which there are fixed costs for offering a product, shelf constraints, or substitution costs. Second, we study the joint assortment and pricing problem, in which the goal is to simultaneously select the set of offered products as well as their prices. We solve the pricing problem optimally when customers have some universal ranking of the products, and hence the tree takes the form of a single path. We also solve the problem optimally on the general tree when the prices are restricted to be quality consistent; higher-quality products must be priced above lower-quality products. Finally, we present computational experiments on both synthetic data and real hotel purchase data. Our estimation procedure shows both how to build the tree of products and how to estimate the underlying arrival probabilities of each customer type from historical sales data. These experiments show that the tree choice model captures customer purchasing behavior more accurately than the multinomial logit choice model in the majority of test cases. The online appendix is available at https://doi.org/10.1287/msom.2017.0662 .

Two-stage supply chain design with safety stock placement decisions

In this paper, we propose a supply chain design model that integrates facility location with safety stock placement and delivery strategy decisions, to reflect their interdependence and ultimately improve the resulting supply chain design. Safety stock placement decisions are modelled using the guaranteed-service approach. We also consider two customer classes that differ with respect to their delivery time preferences. The resulting non-linear model is formulated as a conic quadratic mixed-integer program that can be solved to optimality. We show that the model can be extended to include demand correlation and stochastic lead times. The computational experiments capture the trade-offs between location and safety stock placement decisions, between demand variability pooling and proximity to retailers, and between lead times and service times. We show that, when the distribution centres have short lead times to the retailers, distribution centres tend to hold safety stocks, and the retailers offer express deliveries. Conversely, when lead times are long, only the retailers hold safety stocks and they do not offer express deliveries.

Analysis of a two-class single-server discrete-time FCFS queue: the effect of interclass correlation

In this paper, we study a discrete-time queueing system with one server and two classes of customers. Customers enter the system according to a general independent arrival process. The classes of consecutive customers, however, are correlated in a Markovian way. The system uses a “global FCFS” service discipline, i.e., all arriving customers are accommodated in one single FCFS queue, regardless of their classes. The service-time distribution of the customers is general but class-dependent, and therefore, the exact order in which the customers of both classes succeed each other in the arrival stream is important, which is reflected by the complexity of the system content and waiting time analysis presented in this paper. In particular, a detailed waiting time analysis of this kind of multi-class system has not yet been published, and is considered to be one of the main novelties by the authors. In addition to that, a major aim of the paper is to estimate the impact of interclass correlation in the arrival stream on the total number of customers in the system, and the customer delay. The results reveal that the system can exhibit two different classes of stochastic equilibrium: a “strong” equilibrium where both customer classes give rise to stable behavior individually, and a “compensated” equilibrium where one customer type creates overload.

Analysis of Networks of the Resource Queuing Systems

Consideration was given to the model of a queuing network with resource queuing systems—the multi-server systems with losses where servicing of the accepted customer occupies random volumes of resources with the given distribution function depending on the customer class and the type of required servicing. Customer servicing in a node can be interrupted by a signal arriving at an exponential time from the start of servicing, and during the time of customer sojourn in the network it may be interrupted more than once. Analytical formulas for calculation of the fundamental probability-time model characteristics—including the joint distribution functions of the number of customers in the nodes and the volumes of resources occupied by them—were proposed on the assumption of the Poisson flows incoming to the nodes and arbitrary servicing.

Prediksi Customer Churn dengan Algoritma Decision Tree C4.5 Berdasarkan Segmentasi Pelanggan untuk Mempertahankan Pelanggan pada Perusahaan Retail

Customer is a very important asset for retail companies. This is the reason why retail companies should plan and use a fairly clear strategy in treating customers. With the large number of customers, the problem that must be faced is how to identify the characteristics of all customers and able to retain existing customers in order not to stop buying and moving to a competitor retail company. By applying the concept of CRM, a company can identify customers by segmenting customers while also being able to implement customer retention programs by predicting potential churn on each customer class. The data used comes from UD.Mawar Sari. Customer segmentation process uses RFM model to get customer class. UD. Mawar Sari customer class is dormant, everyday, golden and superstar. The construction of prediction models using the Decision Tree C4.5. The application of the prediction model obtains performance results, that is: Dormant: Recall 97.51%, Precision 75.18%, Accuracy 76.18%. Everyday: Recall 100%, Precision 99.04%, Accuracy 99.04%. Golden: Recall 100%, Precision 98.84%, Accuracy 98.84%. Superstar: Recall 96.15%, Precision 99.43%, Accuracy 95.63%. Results of the evaluation with confusion matrix it can be concluded that the dormant customer class is a potentially churn customer class.

Calculation of the performance region of an easy-to-optimize alternative for Generalized Processor Sharing

Service differentiation is a basic requirement in every modern queueing system with multiple classes of customers. In this paper, we look at Hierarchical Generalized Processor Sharing (H-GPS), which is a discrete-time hierarchically-structured implementation of the well-known idealized Generalized Processor Sharing (GPS) scheduling discipline. We prove that, for three classes, H-GPS can be configured to obtain any performance possible by other scheduling mechanisms, such as priority queueing or GPS. The hierarchical nature of a H-GPS system, however, has the major advantage that optimization is easier and more intuitive. To this end, we also present an algorithm to calculate the configuration parameters for H-GPS given a certain performance objective.

Price elasticities of retail energy demands in the United States: New evidence from a panel of monthly data for 2001–2016

Price elasticities play an important role in energy demand forecasting, which in turn shapes energy policy and investment decisions. However, there is still considerable debate around how responsive customers are to energy prices, and whether investments in metering and consumer education have made them more responsive in recent decades. Using a Generalized Leontief (GL) demand system and a rich panel of monthly data covering 2001–2016 for the lower 48 United States, we estimate own- and cross-price elasticities of retail demand for electricity, natural gas and fuel oil for the three major customer classes: residential, commercial and industrial. These estimates indicate that retail energy demand in the United States was highly price-inelastic over 2001–2016, consistent with historical and current estimates used by many utility planners. Our findings suggest that, with current technologies and behavior, higher energy prices will not induce significant reductions in demand. Hence, energy efficiency standards and utility demand-side programs are still an important strategy for managing energy demand growth, mitigating energy price risk, and reducing the environmental impacts associated with energy use. Finally, while our analysis uses data from the United States, our approach is general and can be readily extended to other countries that have similar data available.

ANALYSIS OF THE NETWORK WITH MULTIPLE CLASSES OF POSITIVE AND NEGATIVE CUSTOMERS AT A TRANSIENT REGIME

This paper is devoted to the investigation of the G-network with multiple classes of positive and negative customers. The purpose of the investigation is to analyze such a network at a transient regime, to find the state probabilities of the network that depend on time. In the first part, a description of the functioning of G-networks with positive and negative customers is provided, when a negative customer when arriving to the system destroys a positive customer of its class. Streams of positive and negative customers arriving at each of the network systems are independent. Services of positive customers of all types occur in accordance with a random selection of them for service. For nonstationary probabilities of network states, a system of Kolmogorov's difference-differential equations (DDE) has been derived. A method for their finding is proposed. It is based on the use of a modified method of successive approximations, combined with the method of series. It is proved that successive approximations converge with time to a stationary probability distribution, the form of which is indicated in this paper, and the sequence of approximations converges to the unique solution of the DDE system. Any successive approximation is representable in the form of a convergent power series with an infinite radius of convergence, the coefficients of which satisfy recurrence relations, which is convenient for computer calculations. A model example illustrating the determination of the time-dependent probabilities of network states using this technique has been calculated. The obtained results can be applied in modeling the behavior of computer viruses and attacks in information and telecommunication systems and networks, for example, as a model of the impact of some file viruses on server resources. variable.

A base-stock inventory model with service differentiation and response time guarantees

In response to customer specific service time guarantee requirements, service providers can offer differentiated services. Although response time agreements offer more flexibility, most of the theoretical models for customer differentiation are based on fill rates, due to the difficulty of calculating the response time distribution in the presence of priorities. In this paper we focus on the (S−1,S,K) model with two customer classes, in which low priority customers are served only if the inventory level is above K. For this model, we derive the exact distribution of the response time (within lead time) for the lower priority class and provide a new closed-form approximation for the fill rate for high priority customers. Via numerical experiments, we show that the fill rate approximation performs comparable to the best approximations in the literature, while its implementation requires only elementary operations.

FINDING EXPECTED REVENUES IN G-NETWORK WITH MULTIPLE CLASSES OF POSITIVE AND NEGATIVE CUSTOMERS

Investigation of the G-network with multiple classes of positive and negative customers has been carried out in the article. The purpose of the investigation is to analyze such a network at a transient regime, finding expected revenues in the network systems depending on time. A negative customer arriving to the system and destroys a positive customer of its class. Streams of positive and negative customers arriving to each of the network systems are independent. Services of positive customers of all types occur in accordance with a random selection of them for service. For the expected revenues, a system of Kolmogorov's difference-differential equations has been derived. A method for their finding is proposed. It is based on the use of a modified method of successive approximations, combined with the method of series. A model example illustrating the finding of time-dependent expected revenues of network systems has been calculated, which shows that the expected revenues of network systems can be either increasing or decreasing time functions. The obtained results can be applied in forecasting losses in information and telecommunication systems and networks from the penetration of computer viruses into it and conducting computer attacks.

Infinite-Horizon Average Optimality of the N-Network in the Halfin–Whitt Regime

We study the infinite-horizon optimal control problem for N-network queueing systems, which consists of two customer classes and two server pools, under average (ergodic) criteria in the Halfin–Whitt regime. We consider three control objectives: (1) minimizing the queueing (and idleness) cost, (2) minimizing the queueing cost while imposing a constraint on idleness at each server pool, and (3) minimizing the queueing cost while requiring fairness on idleness. The running costs can be any nonnegative convex functions having at most polynomial growth. For all three problems, we establish asymptotic optimality; namely, the convergence of the value functions of the diffusion-scaled state process to the corresponding values of the controlled diffusion limit. We also present a simple state-dependent priority scheduling policy under which the diffusion-scaled state process is geometrically ergodic in the Halfin–Whitt regime, and some results on convergence of mean empirical measures, which facilitate the proofs.

Pseudo conservation for partially fluid, partially lossy queueing systems

We consider a queueing system with heterogeneous customers. One class of customers is eager; these customers are impatient and leave the system if service does not commence immediately upon arrival. Customers of the second class are tolerant; these customers have larger service requirements and can wait for service. In this paper, we establish pseudo-conservation laws relating the performance of the eager class (measured in terms of the long run fraction of customers blocked) and the tolerant class (measured in terms of the steady state performance, e.g., sojourn time, number in the system, workload) under a certain partial fluid limit. This fluid limit involves scaling the arrival rate as well as the service rate of the eager class proportionately to infinity, such that the offered load corresponding to the eager class remains constant. The workload of the tolerant class remains unscaled. Interestingly, our pseudo-conservation laws hold for a broad class of admission control and scheduling policies. This means that under the aforementioned fluid limit, the performance of the tolerant class depends only on the blocking probability of the eager class, and not on the specific admission control policy that produced that blocking probability. Our psuedo-conservation laws also characterize the achievable region for our system, which captures the space of feasible tradeoffs between the performance experienced by the two classes. We also provide two families of complete scheduling policies, which span the achievable region over their parameter space. Finally, we show that our pseudo-conservation laws also apply in several scenarios where eager customers have a limited waiting area and exhibit balking and/or reneging behaviour.

Expanding Women’s Access To The Banking Sector

The increase in micro, small, and medium enterprise (UMKM) units managed by women has contributed a positive impact on reducing poverty rates. The data highlighting lack of female entrepreneurs of UMKM who used and/or are supported with financing facility from banks have led to some emerging questions to be discussed. This present study promotes significance, namely: 1) The effectiveness of financing tends to affect the banks’ performance through created demands; 2) The results of this study are expected to enhance the attention of Islamic banking to female customers by increasing the allocation of financing for those customers; 3) The further implication of these findings is the increase in national economic growth from the UMKM actuated by women; 4) The broader implication is the reduction of working women at high risks because of the conducive business climate in Indonesia with the support of the banking sector. Therefore, this quantitative research that was conducted with regression analysis focused on the internal factors to analyze perceptions, behavior, culture and social class of female customers towards the effectiveness of working capital financing at Islamic banking in Palembang. The findings revealed that the variable of perceptions gave a positive effect on the effectiveness of financing of 0.316, behavior showed a positive impact on the effectiveness of financing of 0.676, culture noted a positive influence on the effectiveness of financing of 0.346, whereas social class demonstrated a negative impact on the effectiveness of the financing of -0.14.

Optimal resource usage in a multiclass customer-based cloud centers

Cloud computing is an evolutionary computing paradigm, whereby shared resources are provided to users on-demand over the internet as services. Although the benefits and opportunities of cloud computing are tremendous, its challenges and problems require a huge amount of effort to be addressed. The ability to deliver guaranteed quality of services is crucial for the commercial success of this computing paradigm. We consider the cloud centers which handle the requests from different classes of customers such as high-priority, low-priority, etc. on the basis of service level agreements. Both the modes have equal chances of receiving service, and one and more virtual machines are used to serve each arrival mode. We analyse the system considering an infinite-buffer multi-server queuing system with two priority classes of impatient customers. We obtain various performance measures to analyse such a system which is too complex due to abandonment phenomenon.

Optimal Resource Usage in a Multiclass Customer Based Cloud Centers

Cloud computing is an evolutionary computing paradigm, whereby shared resources are provided to users on-demand over the internet as services. Although the benefits and opportunities of cloud computing are tremendous, its challenges and problems require a huge amount of effort to be addressed. The ability to deliver guaranteed quality of services is crucial for the commercial success of this computing paradigm. We consider the cloud centers which handle the requests from different classes of customers such as high-priority, low-priority, etc. on the basis of service level agreements. Both the modes have equal chances of receiving service, and one and more virtual machines are used to serve each arrival mode. We analyse the system considering an infinite-buffer multi-server queuing system with two priority classes of impatient customers. We obtain various performance measures to analyse such a system which is too complex due to abandonment phenomenon.

Times until service completion and abandonment in an M/M/<inline-formula><tex-math id="M1">\begin{document}$ m$\end{document}</tex-math></inline-formula> preemptive-resume LCFS queue with impatient customers

We consider an M/M/$ m$ preemptive-resume last-come first-served (PR-LCFS) queue without exogenous priority classes of impatient customers. We focus on analyzing the time interval from the arrival to either service completion or abandonment for an arbitrary customer. We formulate the problem as a one-dimensional birth-and-death process with two absorbing states, and consider the first passage times in this process. We give explicit expressions for the probabilities of service completion and abandonment. Furthermore, we present sets of recursive computational formulas for calculating the mean and second moment of the times until service completion and abandonment. The two special cases of a preemptive-loss system and an ordinary M/M/$ m$ queue with patient customers only, both incorporating the preemptive LCFS discipline, are treated separately. We show some numerical examples in order to demonstrate the computation of theoretical formulas.

A Continuous-Class Queueing Model With Proportional Hazards-Based Routing

Motivated by jail overcrowding and the U.S. correctional system's widespread use of risk models to aid in inmate release decisions both prior to trial (i.e., pretrial release) and near the end of their jail sentence (i.e., split sentencing), we formulate and analyze a queueing network model with two novel features: there is a continuum of customer classes corresponding to an inmate's continuous risk level p, and routing in the network is dictated by a Cox proportional hazards model, where the hazard rate associated with recidivism (i.e., committing another crime) during release is proportional to eγp for some parameter γ. We perform an exact analysis of a continuous-class M/M/c/c (i.e., Erlang B) model where preemptive priority is awarded according to the risk level p, and use it to develop approximate performance measures for a queueing network of a jail with a two-threshold policy, which dictates who is granted pretrial release and who receives a split sentence. For a slightly simplified version of the model, we derive sufficient conditions under which no inmates are offered pretrial release unless all inmates are given a split sentence.

Smart Meter Data-Based Load Profiles and Their Effect on Distribution System State Estimation Accuracy

The constantly spreading smart meters open new possibilities for load profiling. With smart meter measurements, standard customer class load profiles can be updated, customers can be clustered into similarly behaving groups, or individual load profiles can be created. This paper studies how these new smart meter data -based load profiling methods affect the distribution system state estimation accuracy. Extensive simulations were made with a real distribution network containing both medium and low voltage networks. Results achieved with different load profiling methods were compared with each other, with publicly available standard customer class load profiles, and with other currently used load modelling methods. In all studied cases, the new smart meter data -based load profiles provided superior state estimation accuracy. The results presented in this paper should motivate distribution network operators to utilize smart meter measurements more efficiently in load profiling and state estimation.

Consumption effects of an electricity decarbonization policy: Hong Kong

This paper estimates the consumption effects of an electricity rate increase triggered by an electricity decarbonization policy's implementation. Underscoring its real-world relevance is the policy's net impact on CO2 emissions, the sum of (a) the supply-side impact attributable to using generation resources with low emissions to displace those with high emissions, and (b) the demand-side impact caused by energy consumption changes in response to the electricity rate increase. For Hong Kong, the changes in (b) are decreases in electricity consumption and increases in town gas consumption. Using a sample of monthly data for the period of 1981–2016, we document the low price responsiveness of Hong Kong's electricity and town gas demands by customer class (residential, commercial and industrial). Hence, the 40% projected electricity rate increase due to Hong Kong's adopted electricity decarbonization policy may only have a small demand-side impact on CO2 emissions. Finally, the electricity demands' low price responsiveness has two important policy implications. First, Hong Kong's demand-side-management should rely more on energy-efficiency improvements than price-induced consumption reductions. Second, restructuring Hong Kong's electricity industry to introduce wholesale competition should consider the potential for large electricity price spikes and market power abuse in connection to price-inelastic electricity demands.

An Ergodic Control Problem for Many-Server Multiclass Queueing Systems with Cross-Trained Servers

A Markovian queueing network is considered with d independent customer classes and d server pools in Halfin–Whitt regime. Class i customers has priority for service in pool i for i = 1, …, d, and may access some other pool if the pool has an idle server and all the servers in pool i are busy. We formulate an ergodic control problem where the running cost is given by a non-negative convex function with polynomial growth. We show that the limiting controlled diffusion is modelled by an action space which depends on the state variable. We provide a complete analysis for the limiting ergodic control problem and establish asymptotic convergence of the value functions for the queueing model.