WSC 2008 Final Abstracts

Nonlinear Process Modeling and Optimization Based on Multiway Kernel Partial Least Squares Model
Liqing Di, Zhihua Xiong, and Xianhui Yang (University of Tsinghua)

Abstract:
MKPLS (Multiway Kernel Partial Least Squares) methods are used to model the batch processes from process operational data. To improve the optimization performance, a batch-to-batch optimization strategy is proposed based on the idea of the similarity between the iterations during numerical optimization and successive batch runs. SQP (Sequential Quadratic Programming) coupling with MKPLS model is used to solve the optimization problem, and the plant data, instead of the MKPLS model predictions, are used in gradient calculation. The proposed strategy is illustrated on a simulated bulk polymerization of styrene. The results demonstrate that the optimization performance has been improved in spite of the model-plant mismatches

Speeding up Call Center Simulation and Optimization by Markov Chain Uniformization
Eric Buist, Wyean Chan, and Pierre L'Ecuyer (Université de Montréal)

Abstract:
Staffing and scheduling optimization in large multiskill call centers is time-consuming, mainly because it requires lengthy simulations to evaluate performance measures. Simplified models that provide tractable formulas are unrealistic in general. In this paper we explore an intermediate solution, based on an approximate continuous-time Markov chain model of the call center. This model is more accurate than the commonly used approximations, and yet can be simulated faster than a more realistic simulation. To speed up the simulation, we uniformize the Markov chain and simulate only its discrete-time version. We show how performance measures such as the fraction of calls answered within a given waiting time limit can be recovered from this simulation, how to synchronize common random numbers in this setting, and how to use this in the first phase of an optimization algorithm based on the cutting plane method. We also discuss various implementation issues and provide empirical results.

A New Policy for the Service Request Assignment Problem
Anshul Sheopuri, Sai Zeng, and Chitra Dorai (IBM T J Watson Research Center)

Abstract:
We study the problem of assigning multiple severity level service requests to agents in an agent pool. Each severity level is associated with a due date and a penalty, which is incurred if the service request is not resolved by the due date. Motivated by Van Meighem (2003), who shows the asymptotic optimality of the Generalized Longest Queue policy for the problem of minimizing the due date dependent expected delay costs when there is a single agent, we develop a class of Index-based policies that is a generalization of the Priority First-Come-First-Serve, Weighted Shortest Expected Processing Time and Generalized Longest Queue policy. In our simulation study of an assignment system of a large technology firm, the Index-based policy shows an improvement of 0-20 % over the Priority First-Come-First-Serve policy depending upon the load conditions.

Business Process Based Simulation: A Powerful Tool for Demand Analysis of Business Process Reengineering and Information System Implementation
Linlin Cui, Yueting Chai, and Yi Liu (Tsinghua Univ.)

Abstract:
Demand analysis is of fundamentally importance in the implementation of information system. Business process reengineering (BPR) often gets involved in the process of demand analysis and play a crucial role in the achievement of project objectives. Business process based simulation (BPS) provides a precise and visual method to analyze and compare the concerned performances before and after BPR. The paper presents an industrial experience in using the BPS tool to demonstrate the effects of BPR on restraining stocking-up and overdue payments in the distribution management of a supply chain. Before significant investment involved, the related design result of BPR is validated both by the analytical method and simulation experiments. Based on the mutual supportive results, the BPS method approves its correctness and show its nicety, flexibility and the capacity of visualization.

Industrial Enterprises Business Processes Simulation with BPsim.MAS
Konstantin Aksyonov (Urals State Technical University - UPI), Elena Smoliy and Alexey Khrenov (NPP Business Support Systems) and Evgeny Bykov (Urals State Technical University - UPI)

Abstract:
Necessity to mind a huge amount of factors while developing modern enterprises models dictates new requirements for modeling software, which needs to process all data to achieve precise results, make use of all possible means for that, such as distributed calculations and introduce original approaches where possible to avoid extra time waste for multiple simulation experiments. The article focuses on the software apparatus used in distributed multi-agent resource conversion processes based tool BPsim.MAS, pointing out its advantages and describing used technologies. Second part describes BPsim.MAS deployment in Urals Industrial Group, CJSC, which allowed income and market share growth. Various pricing strategies are discussed and active/passive competitors’ behavior is considered.

Simulation of Process Execution Monitoring and Adjustment Schemes
Russell R. Barton and Jun Shu (The Pennsylvania State University)

Abstract:
Optimization and design of production and service operations has been a cornerstone of simulation applications for many years. Recently there has been increasing interest in excellence in process execution, for service processes, supply chains, and manufacturing operations. Simulation again has an important role to play, in evaluating the effectiveness of different execution monitoring schemes. The focus of analysis changes in this setting, however, from monitoring characteristics of resources, such as queue length, waiting time and utilization, to monitoring the timeliness and correctness of the progress of entities through the system, and reacting with realtime adjustments to resources and routings. This paper describes the types of simulation analysis needed, and identifies difficulties in conducting such simulation studies using current commercial software.

Towards a Flexible Business Process Modeling and Simulation Environment
Changrui Ren, Wei Wang, Jin Dong, Hongwei Ding, Qinhua Wang, and Bing Shao (IBM China Research Laboratory)

Abstract:
Business process is crucial to the success of any business. Business process modeling enables a common understanding and analysis of a business process, and simulation is an effective way to diagnose and evaluate complex business processes. There are lots of software tools in market for business process modeling and simulation, however, a common issue for these tools is the conflict between usability and flexibility. The improvement of usability often means the reduction of flexibility. This paper introduces an IBM asset named Supply Chain Process Modeler (SCPM), which aims at providing a tailored business process modeling and simulation environment for business users. SCPM tries to achieve a better trade-off between usability and flexibility with two major efforts, a process repository for process modeling, and a flexible environment for process simulation. A case study from real consulting practice is also included to illustrate how SCPM works for business process simulation.

Simulation Modeling and Analysis for In-Store Merchandizing of Retail Stores with Enhanced Information Technology
Kanna Miwa (Nagoya Gakuin University) and Soemon Takakuwa (Nagoya University)

Abstract:
In this study a procedure of simulation modeling for in-store merchandizing was used in order to examine customer flows in retail stores. One of the main objectives of this paper is to propose an approach of building a simulation model by making use of the POS data. The methods for identifying the arrival time and routing inside store are described, using daily POS data. This study primarily analyzed customer congestion and the utilization of the cash registers. Recently, IC tags are recognized in the fields of manufacturing and distribution systems as a powerful tool to reduce sales transaction time, with the assistance of the POS system. The proposed procedure was applied to the actual case. The results show that the customer flow can be examined easily, while the proposed procedure is useful for store management including cashiers scheduling and in-store merchandising.

Simulation Based Sales Forecasting on Retail Small Stores
Hairong Lv, Xinxin Bai, Wenjun Yin, and Jin Dong (IBM China Research Laboratory)

Abstract:
As the competition becomes more and more intense, many retail small store chain operators are eager to know how to evaluate new store locations quantitatively to support a scientific business development decision, e.g. what will be the potential sales of new stores? How will a new store influence other existing stores? In this paper, we propose a novel framework that can predict the competitions between new store and competing stores. First, we examine the special characteristics of customers' demand on retail small stores and develop several models to estimate different customers’ demand; and then model the relationship between customers and multiple stores, such as route model, user behavior model; finally, we compute the store sales via simulation. A real case on the micro-competition analysis in a China city for one of the famous retail small store chain operators shows that our method is more practical and accurate than other methods.