In call centers, the service level is an important metric to measure the reasonability of the staffing schedule. Traditional service level calculation methods are based on the queue theory, which has very strict restrictions and is not suitable for real scenarios. Therefore, in this paper, a data-driven method to solve the service level prediction problem is proposed to be used. To this end, the relationship between service level and other factors, such as number of calls, number of agents, time, is explored. Then some features are extracted based on empirical analyses and propose to use decision tree based ensemble methods, like random forest and GBDT, to model the relationship between service level and input features. Finally, extensive experimental results show that the proposed method outperforms other baselines significantly. Especially compared with the traditional queue theory methods, our method improves the performance by 6% and 9% in terms of MAE and MAPE.

Process recommendation is an essential technique to help process modeler effectively and efficiently model a manufacturing process from scratch. However, the current process recommendation methods suffer from the following problems: (1) To extract all the execution paths from a manufacturing process, the behavior-based methods may occur a state space explosion problem when unfolding a process with multiple parallel patterns, resulting in low efficiency. (2) Current structure-based methods are inefficient since too many expensive computations of the graph edit distance are involved. (3) Most of the existing methods manually design their process similarity metrics with several features, which can only be applied in specific situations. (4) Few works provide visualization tools for process modeling assistance. To resolve these problems, this paper proposes a context-aware recommendation system for improving manufacturing process modeling. First, the independent paths and P,Q-grams are efficiently extracted from the manufacturing processes in the repository to represent their typical behavior and structure. Then, the process recommendation problem is transformed into the word prediction problem in natural language processing, where the serialization of an independent path/P,Q-gram and a node in it are separately regarded as a sentence and a word. The Word2vec model is introduced to automatically learn the relationships among nodes from independent paths and P,Q-grams and generate the vectors with hundreds of context-aware features for nodes in the repository. After that, the top-k similar nodes are recommended for the target node in the process fragment under construction based on the k-nearest neighbors algorithm. Finally, a visualization tool is provided for process modelers to efficiently design a new manufacturing process. Experimental evaluations show that the proposed method can perform similar or even better than the baseline methods in terms of recommending quality.

The development of mobile workflow management systems (mWfMS) leads to large number of business process models. In the meantime, the location restriction embedded in mWfMS may result in different process models for a single business process. In order to help users quickly locate the difference and rebuild the process model, detecting the difference between different process models is needed. Existing detection methods either provide a dissimilarity value to represent the difference or use predefined difference template to generate the result, which cannot reflect the entire composition of the difference. Hence, in this paper, we present a new approach to solve this problem. Firstly, we parse the process models to their corresponding refined process structure trees (PSTs), that is, decomposing a process model into a hierarchy of subprocess models. Then we design a method to convert the PST to its corresponding task based process structure tree (TPST). As a consequence, the problem of detecting difference between two process models is transformed to detect difference between their corresponding TPSTs. Finally, we obtain the difference between two TPSTs based on the divide and conquer strategy, where the difference is described by an edit script and we make the cost of the edit script close to minimum. The extensive experimental evaluation shows that our method can meet the real requirements in terms of precision and efficiency.

In recent years, many ride-sharing systems have been widely used for helping people travel around. These systems generate a lot of data every day, e.g., the trajectories and the transactions. But as far as we know, there is no such monitoring system or platform for ride sharing systems to manage and observe this set of data in an effective way. Hence, in this paper, we illustrate YoRoad, a general monitoring service for ride-sharing system. YoRoad consists of three main modules: (1) moving objects tracking module is used for monitoring the drivers real-time locations, where moving object indexing and spatial query processing technique is embedded; (2) user management module, where the information for each driver and rider is maintained; and (3) data statistics module, where we perform some data analysis such as the total distance saving comparison between ride-sharing and non-ride-sharing, trajectory data mining and location preference or travel pattern discovery based on the transaction data.

This paper presents a system that utilizes process recommendation technology to help design new business processes from scratch in an efficient and accurate way. The proposed system consists of two phases: 1) offline mining and 2) online recommendation. At the first phase, it mines relations among activity nodes from existing processes in repository, and then stores the extracted relations as patterns in a database. At the second phase, it compares the new process under construction with the premined patterns, and recommends proper activity nodes of the most matching patterns to help build a new process. Specifically, there are three different online recommendation strategies in this system. Experiments on both real and synthetic datasets are conducted to compare the proposed approaches with the other state-of-the-art ones, and the results show that the proposed approaches outperform them in terms of accuracy and efficiency.

Production systems based on knowledge rules have been widely used for reasoning both in industry and academia. However, rule matching in production system is time-consuming too much and it always incur the system crash when the massive knowledge exceeds the limitations of memory and computing capacity of one single computer. The advent of cloud computing—a new on-demand computing model brings us an inspiring perspective to address this problem. In this paper, a MapReduce-based rule matching method was proposed. It decomposes the task of rule matching and maps subtasks to different computers in a distributed and parallel computing environment, and gets the final matching result after reduce phase. An experimental evaluation shows the high efficiency of the method.

Operations management is important to a company, so more and more business process models are created. At the same time, how to manage such a large amount of process models is becoming a big challenge for companies. Querying the relevant process models is proposed as a business process management technology and it has attracted more and more attention by researchers. The existing methods query the relevant models for a query process model by measuring their similarities. And most of them measure the similarity by focusing on only one kind of feature, such as the structural features or behavioral features, while ignoring other features. In this paper, we consider both structural features and behavioral features to query the relevant process models for a query process model. In order to reach this goal, we use two classification methods named back propagation neural network (BPNN) and support vector machines(SVM) for classifying the candidate models in the repository into two classes: relevant and irrelevant. For the sake of classification, we summarize 7 features to represent the similar or dissimilar parts of two process models. The experiment result shows the precision and efficiency of the classification methods are acceptable.

Recently, many ride sharing systems have been commercially introduced (e.g., Uber, Flinc, and Lyft) forming a multi-billion dollars industry. The main idea is to match people requesting a certain ride to other people who are acting as drivers on their own spare time. The matching algorithm run by these services is very simple and ignores a wide sector of users who can be exploited to maximize the benefits of these services. In this demo, we demonstrate SHAREK; a driver-rider matching algorithm that can be embedded inside existing ride sharing services to enhance the quality of their matching. SHAREK has the potential to boost the performance and widen the user base and applicability of existing ride sharing services. This is mainly because within its matching technique, SHAREK takes into account user preferences in terms of maximum waiting time the rider is willing to have before being picked up as well as the maximum cost that the rider is willing to pay. Then, within its course of execution, SHAREK applies a set of smart filters that enable it to do the matching so efficiently without the need to many expensive shortest path computations.

The increasing number of mobile services makes users confused to select appropriate services among plenty of service icons or links. Current developers always choose to recommend recently or mostly used services to users, but these approaches neglect the relations between user states and environment information and invocations, and the recommendation results will not be accurate when the mobile services are invoked evenly. In this paper, we propose a novel approach to recommend services on mobile devices to user. Firstly, we design a user behavior model by taking advantage of user's mobile context information like time and location to describe the user states. Secondly, we design a generate model to explain how the sequential service invocations are generated by analyzing the collected sequential history record of mobile users. Thirdly, we adopt logistic model tree approach to determine user state according to given mobile context information, and recommend services to user according to his user state. The experiment results show that our approach performs better than baseline approaches.

Efficient query processing over a large amount of business process models is important for managing the business process model repository. The structural similarity between two process models is considered as the main measurement for ranking the process models for a given search model. Current business process query methods are inefficient since too many expensive computations of the graph edit distance are involved for constructing the elements mapping as well as deriving the structural similarity. To address this, using Petri-net as the modelling method, this paper presents the Hungarian algorithm based query method, where we firstly define the context similarity for a pair of place nodes that are from different process models by taking into account both the common paths and common transitions, then transform the elements (e.g., The transitions and the places) mapping to classical assignment problem that can be solved by Hungarian algorithm efficiently. In this way, we can save a lot of time for searching the best combination of elements mapping. Finally, we use the common method of the graph edit distance to measure the structural similarity based on the found best combination of elements mapping.

Nowadays, many scientific workflows are deployed in the cloud, and how to schedule the tasks according to the users' QoS (Quality of Service) requirements, such as the make span and the monetary cost, has been proposed as the main challenge. In this paper, we aim to solve the problem of finding the scheduling solutions to minimize the workflow make span under the constraint of the user's budget. Considering it is very time consuming to find the optimal solution, instead, we adopt an evolutionary computation technique called Particle Swarm Optimization (PSO) to derive the approximate answers. The proposed method is evaluated with real scientific workflows of different structures and sizes. Comparing with the latest method, the experiment results show that our proposed approach can achieve better performance by increasing the number of particles and iterations.

With increasing improvement of the business process management (BPM) technology, large-scale business process repositories have been adapted widely. However, due to the explosion of the number of business processes, a large part of enterprises are confronting with the challenge on effective management of those massive processes. Usually, each business process is modeled as a process graph, and therefore most existing approaches are based on graph mining algorithms. This paper puts forward a new method, which first utilizes the breadth first search (BFS) algorithm to label the process model, and then calculates the similarity based on the matching distance. The experimental results show that our method is efficient enough for practical use, especially suitable for fuzzy retrieval.

The traditional process recommendation techniques could not handle processes with cycle structures,so the practicability was limited greatly.For this problem,a new recommendation method supporting cycle structure was presented,which computed the distance between reference process and process patterns based on maximum common subgraph and minimum common supergraph,and the appropriate candidate nodes selected by computation was used as recommend result.Combined with traditional process recommendation technique,JTangWFR prototype system was implemented.Experiment results with synthetic and real-world datasets proved that the proposed system not only could support cyclic structures effectively,but also have great improvement in efficiency and accuracy of recommendation compared to the traditional techniques.

Graph data has been commonly used and widely researched both in academia and industry for many applications. And measuring similarity between graphs (i.e., graph matching) is the essential step for graph searching, pattern recogni tion and machine vision. At present, the most widely used approach to address the graph matching problem is graph edit distance (GED). However, the computation complexity of GED is expensive and it takes unacceptable time when the graph becomes larger. Generally, graph could be canonical labeled by some sort of strings and we use the depth-first search (DFS) code as our canonical labeling system. Ba sed on DFS codes, combining the Levenshtein distance (i.e., string edit distance, SED), we proposed a novel method for similarity measurement of graphs. Processing and calculating the distance between two DFS codes, we turned the graph matching problem into string matching, which gains great improvement on the matching performance. The experimental results prove its usefulness.

Process retrieval is critical for workflow repository management. Structural similarity metric based on graph matching could achieve highest retrieval quality. Nowadays, researchers mainly adopt graph edit distance (GED) as the approach for comparing process models. However, the computation complexity of GED based methods are high and their cost functions depend heavily on the application domain. To overcome these shortcomings, we use the maximal common subgraph (MCS) approach instead and propose a depth-first search (DFS) code based method to implement the MCS. The minimum DFS codes are used to canonically label the process models and their fragments. By comparing the minimum DFS codes of the fragments, the maximal common subgraphs between the search model (i.e., a given process model or fragment) and the processes in the repository could be found. The experimental evaluations show that our method is feasible for real applications.

We present research aimed at improving our understanding of the use and implementation of quantification in relational query languages in general and SQL in particular. In order to make our results as general as possible, we use the framework of Generalized Quantification. Generalized Quantifiers (GQs) are high-level, declarative logical operators that in the past have been studied from a theoretical perspective. In this paper we focus on their practical use, showing how to incorporate a dynamic set of GQs in relational query languages, how to implement them efficiently and use them in the context of SQL. We present experimental evidence of the performance of the approach, showing that it improves over traditional (relational) approaches.

In modern commerce, both frequent changes of custom demands and the specialization of the business process require the capacity of modeling business processes for enterprises effectively and efficiently. Traditional methods for improving business process modeling, such as workflow mining and process retrieval, still requires much manual work. To address this, based on the structure of a business process, a method called workflow recommendation technique is proposed in this paper to provide process designers with support for automatically constructing the new business process that is under consideration. In this paper, with the help of the minimum depth-first search (DFS) codes of business process graphs, we propose an efficient method for calculating the distance between process fragments and select candidate node sets for recommendation purpose. In addition, a recommendation system for improving the modeling efficiency and accuracy was implemented and its implementation details are discussed. At last, based on both synthetic and real-world datasets, we have conducted experiments to compare the proposed method with other methods and the experiment results proved its effectiveness for practical applications.

The frequent changes of custom demands and the professionalization of the business process both bring huge challenges for the enterprise workflow modeling or design.Traditional methods for improving process design,such as process mining and process retrieval,still involve much manual work.How to greatly improve the efficiency and accuracy for business process design has become a burning problem.The traditional recommendation technique provides us the possible solution.This paper proposes a process recommendation technique which starts from the process repository.Graph mining technique is used to extract the process patterns.Based on near neighbor and maximal subgraph first(NMSF) strategy,we judge the equvalance between the reference process and process patterns for derterming the candidate node sets for recommendation.A prototype called JTangWFR is implemented,and its recommendation scenarios and implementation details are also discussed.The experimental evaluation proves that our system is efficient and effective for practical use.

How to improve the modeling efficiency and accuracy has become a burning problem. The popularization of recommendation technique in E-Commerce provide us new trajectories that can be used for addressing the problem. In this paper, we propose a graph-based workflow recommendation for improving business process modeling. The start point is so-called "workflow repository" including a set of already developed process models. Graph mining method is used to extract the process patterns from the repository. Based on graph edit distance (GED) [2], we calculate the distance between patterns and the partial business process, viewed as reference model, which is under modeling and select the candidate nodes with smaller distances for recommendation. The performance study show its feasibility for practical uses.

Process/workflow mining aims at discovering the underlying processes to help in improving or rebuilding business processes. Most of the current practices of process mining are based on event logs from Transactional Information Systems (TIS) (such as WFM, ERP, CRM, SCM and B2B systems). However, with the popular deployment and use of business rule engine with TIS, a great number of rule logs are generated, but they are rarely utilized for discovering processes. This paper intends to propose a different perspective for process discovery as compared with the traditional way based on the event logs. Firstly, it illustrates a motivation scenario about process mining from rule logs and then brings forward a framework for process discovery based on rule logs. After that, the mining algorithm called Alpha-r with a case study is introduced to discover a process through mining the relations of traces in rule flow log. Finally, some experiments show the effectiveness and performance of the method.