An interactive approach for Bayesian network learning using domain/expert knowledge

Using domain/expert knowledge when learning Bayesian networks from data has been considered a promising idea since the very beginning of the field. However, in most of the previously proposed approaches, human experts do not play an active role in the learning process. Once their knowledge is elicited, they do not participate any more. The interactive approach for integrating domain/expert knowledge we propose in this work aims to be more efficient and effective. In contrast to previous approaches, our method performs an active interaction with the expert in order to guide the search based learning process. This method relies on identifying the edges of the graph structure which are more unreliable considering the information present in the learning data. Another contribution of our approach is the integration of domain/expert knowledge at different stages of the learning process of a Bayesian network: while learning the skeleton and when directing the edges of the directed acyclic graph structure.     

Supporting complex group decisions: A probabilistic multi-dimensional scaling approach

Groups often face complex decisions; decisions in which the decision alternatives are not clearly defined and the criteria for choosing an alternative are subject to dispute within the group. We present a Group Decision Support System that will use judgments from the group to visualize the decision problem in a probabilistic geometric space. In this geometric representation, actual decision alternatives and an ideal alternative—an artificial alternative that identifies the ideal solution to the group's decision dilemma—are portrayed as distributions in a multi-dimensional space. Dispersions of the distributions measure the uncertainties of the decision process. The psychometric theory used to develop the probabilistic geometric representation is described. Preliminary research is presented which demonstrates that geometric representations of this type help groups both to understand better the decision they face and to find better solutions.     

Bayesian network approach to multinomial parameter learning using data and expert judgments

One of the hardest challenges in building a realistic Bayesian Network (BN) model is to construct the node probability tables (NPTs). Even with a fixed predefined model structure and very large amounts of relevant data, machine learning methods do not consistently achieve great accuracy compared to the ground truth when learning the NPT entries (parameters). Hence, it is widely believed that incorporating expert judgments can improve the learning process. We present a multinomial parameter learning method, which can easily incorporate both expert judgments and data during the parameter learning process. This method uses an auxiliary BN model to learn the parameters of a given BN. The auxiliary BN contains continuous variables and the parameter estimation amounts to updating these variables using an iterative discretization technique. The expert judgments are provided in the form of constraints on parameters divided into two categories: linear inequality constraints and approximate equality constraints. The method is evaluated with experiments based on a number of well-known sample BN models (such as Asia, Alarm and Hailfinder) as well as a real-world software defects prediction BN model. Empirically, the new method achieves much greater learning accuracy (compared to both state-of-the-art machine learning techniques and directly competing methods) with much less data. For example, in the software defects BN for a sample size of 20 (which would be considered difficult to collect in practice) when a small number of real expert constraints are provided, our method achieves a level of accuracy in parameter estimation that can only be matched by other methods with much larger sample sizes (320 samples required for the standard machine learning method, and 105 for the directly competing method with constraints).     

Social network topology: a Bayesian approach

Elucidating the pattern of links within social networks is a challenging problem. Of particular difficulty is determining the existence of links in those groups that take active measures to conceal their internal connections, such as terrorist or criminal organizations where conventional social network analysis data-gathering techniques cannot be applied. Network representations of such organizations are useful, because they often represent a useful point of departure in thinking both about the potential capabilities of organizations and how to conduct effective measures to counter them. Developing an effective process for constructing such network representations from incomplete and limited data of variable quality is a topic of much current interest. Here, a method based on Bayesian inference is presented that probabilistically infers the existence of links within a social network. It is tested on data from open source publications. Additionally, the method represents a possible approach to dynamically modelling networks, as it is feasible to calculate how a network will reconfigure following an intervention.     

An efficient heuristic approach for a multi-period logistics network redesign problem

In this paper, a multi-period logistics network redesign problem arising in the context of strategic supply chain planning is studied. Several aspects of practical relevance are captured, namely, multiple echelons with different types of facilities, product flows between facilities in the same echelon, direct shipments to customers, and facility relocation. A two-phase heuristic approach is proposed to obtain high-quality feasible solutions to the problem, which is initially modeled as a large-scale mixed-integer linear program. In the first phase of the heuristic, a linear programming rounding strategy is applied to find initial values for the binary location variables. The second phase of the heuristic uses local search to correct the initial variable choices when a feasible solution is not identified, or to improve the initial feasible solution when its quality does not meet given criteria. The results of a computational study are reported for randomly generated instances comprising a variety of logistics networks.     

Valuation of urban industrial land: An analytic network process approach

In this paper an application of the Analytic Network Process (ANP) to asset valuation is presented. It has two purposes: solving some of the drawbacks found in classical asset valuation methods and broadening the scope of current approaches. The ANP is a method based on Multiple Criteria Decision Analysis (MCDA) that accurately models complex environments. This approach is particularly useful in problems which work with partially available data, qualitative variables and influences among the variables, which are very common situations in the valuation context. As an illustration, the new approach has been applied to a real case study of an industrial park located in Valencia (Spain) using three different models. The results confirm the validity of the methodology and show that the more information is incorporated into the model, the more accurate the solution will be, so the presented methodology stands out as a good alternative to current valuation approaches.     

Supporting rural telecommunications: A compromise solutions approach

Rural telecommunication networks can foster the development of needed infrastructures for rural residents in an economically viable way. Thanks to emerging broadband fiber optics technology, rural telecommunications can be established via hub cities that function as service centers for neighboring smaller rural areas. Determining hub locations typically requires trade-offs among conflicting criteria. Policy makers typically set their goals as target values. The main aim of this paper is to present a zero—one compromise programming model that reflects such policy makers' target-setting behavior under a multiple criteria environment. A case in which the model is applied to the location of hub telecommunication centers in an American state further illustrates the applicability of the method.     

Multiprocessor task scheduling in multistage hybrid flow-shops: a genetic algorithm approach

This paper considers multiprocessor task scheduling in a multistage hybrid flow-shop environment. The objective is to minimize the make-span, that is, the completion time of all the tasks in the last stage. This problem is of practical interest in the textile and process industries. A genetic algorithm (GA) is developed to solve the problem. The GA is tested against a lower bound from the literature as well as against heuristic rules on a test bed comprising 400 problems with up to 100 jobs, 10 stages, and with up to five processors on each stage. For small problems, solutions found by the GA are compared to optimal solutions, which are obtained by total enumeration. For larger problems, optimum solutions are estimated by a statistical prediction technique. Computational results show that the GA is both effective and efficient for the current problem. Test problems are provided in a web site at www.benchmark.ibu.edu.tr/mpt-hfsp.     

Choosing a leader on a complex network

In many real life applications a group of people interact through a communication network, mathematically modelled as a connected graph linking each element of the group. These participants may have diverse objectives and play very different roles depending on their knowledge and privileges. We focus on a particular scenario, in which a certain node is absolutely essential for completing the intended task. Moreover, if a technical failure results in disconnection of a participant to this leader node, this participant can no longer take part in the group's performance.     

Inverse problems of anomalous diffusion theory: An artificial neural network approach

The results are presented of computer simulation of the operation of a three-layer perceptron trained for solving inverse problems of anomalous diffusion theory. Several types of inverse problems are considered, including the problem of determining the Hurst exponent of a selfsimilar medium.     

Identifying systemically important financial institutions: a network approach

Computational Management Science - The Basel Committee on Banking Supervision has proposed a methodology to identify Systemically Important Financial Institutions based on a series of indicators...     

The competitive salesman problem on a network: a worst-case approach

We provide a complexity analysis of the problem of optimal routing of a server on a transportation network in the presence of a competing server. The server that reaches a node first gets the profit from the node. The objective is to maximize the worst-case profit.     

An object-oriented neural network approach to short-term traffic forecasting

This paper discusses an object-oriented neural network model that was developed for predicting short-term traffic conditions on a section of the Pacific Highway between Brisbane and the Gold Coast in Queensland, Australia. The feasibility of this approach is demonstrated through a time-lag recurrent network (TLRN) which was developed for predicting speed data up to 15 minutes into the future. The results obtained indicate that the TLRN is capable of predicting speed up to 5 minutes into the future with a high degree of accuracy (90–94%). Similar models, which were developed for predicting freeway travel times on the same facility, were successful in predicting travel times up to 15 minutes into the future with a similar degree of accuracy (93–95%). These results represent substantial improvements on conventional model performance and clearly demonstrate the feasibility of using the object-oriented approach for short-term traffic prediction.     

A measure of competitiveness in leagues: a network approach

Many sports are played competitively in a league format. Final positions are based on the aggregations of the points won at each game. Issues of promotion, relegation and much else will depend on the position in the league. However, the results may also be seen to constitute a network of inter-team relations in which the links represent the degree to which a pair of teams have similar performance. This idea is taken as the basis for the construction of a systemic measure of competitiveness in the league. The basis for the model is the construction of a blockmodel on a network of binary relations. The method is illustrated by application to nine seasons of the English soccer Premier League.     

An approach for solving a modification of the extended rapid transit network design problem

In this paper we deal with a slight modification of the extended rapid transit network design problem to allow circular lines. A two-stage approach is proposed for solving this problem. In the first stage, an integer model is solved for selecting the stations to be constructed and the links between them. It drastically reduces the dimension of a modification of a 0–1 model given in the literature to adapt it to our problem. In the second stage, the line design problem is solved by means of a procedure that assigns each selected link to exactly one line under certain constraints. We report some computational experiments that show that our approach also produces a drastic reduction on the computational effort required for solving the modification of the 0–1 model given in the literature.     

The “DUEL” expert system: Implementation of a dual approach for IBM-compatible computers

We describe the “DUEL” expert system, a software package intended for analyzing problem situations and decision making in medical, economic, engineering, and other systems, based on modern information technology, the introduction of knowledge and data bases, and deductive and inductive approaches to the synthesis and justification of decisions. “DUEL” is two inference machines: a deductive machine that uses the products contained in the knowledge base, and an inductive machine that provides generalization on the basis of precedents accumulated in the data base. Translated fromDinamicheskie Sistemy, No. 13, 1994, pp. 93–98.     

Synchronization analysis of a complex network family

This paper is concerned with a meaningful and challenging problem: Is it possible to establish synchronization criteria for a network family, the members of which have different network structure? Based on the parameter-dependent Lyapunov function, some synchronization criteria are proposed to address this problem. These simple yet generic criteria can lead to the advantage of easy synchronization verification of every member from the network family. On the other hand, the problem mentioned above can be regarded as a synchronization problem for a network with structure uncertainty, which is also investigated in this paper. Numerical examples are given to demonstrate the effectiveness of the derived criteria.     

An integrated approach for deriving priorities in analytic network process

A multiple objective programming approach for the analytic network process (ANP) is proposed to obtain all local priorities for crisp or interval judgments at one time, even in an inconsistent situation. The weakness of the ANP and fuzzy ANP (FANP) is that the complexity of generating priorities is equal to the number of comparison matrices. In the proposed approach, all sets of crisp priorities for each pairwise comparison matrix can be obtained directly. Moreover, from the outcomes of three examples, the power to reach a limiting supermatrix is less than or equal to the power of the FANP. Thus, the proposed approach can be regarded as an efficient alternative of the fuzzy ANP.     

An integrated approach to worker assignment,workforce flexibility acquisition,and task rotation

In a manufacturing environment, workforce flexibility can be achieved by cross-training and improved via job rotation. In firms with a flexible workforce, employees perform different tasks and functions in response to fluctuations in both product demands and labour resources. This paper presents a mathematical programming model that assigns workers to tasks, rotates workers between the tasks, and determines the training schedule. The objective is to minimize the total costs including training cost, flexibility cost, and productivity loss cost. A constructive-search heuristic is also developed to solve the proposed model. The algorithm provides good solutions in two phases: construction and improvement. At the construction phase, a solution is built using some problem-specific information. The quality of the solution is then enhanced by changing worker assignments at a particular time point during a planning horizon. Our computational results for a number of randomly generated test problems confirms the efficiently of the proposed method.