Qualitative Constraints Research Articles

Plan decoupling of agents with qualitatively constrained tasks

In many task-planning domains, assemblies of autonomous agents need to construct plans and schedules for executing their assigned sets of tasks in order to complete them. Often, there will be dependencies between tasks to be executed by different agents. In such a situation, a plan-coordination problem arises when the joint plan is required to be feasible, whatever locally-feasible plans the individual agents come up with. This problem can be solved by plan decoupling, which is to add a minimum number of constraints such that each agent can make a plan for its set of tasks independently of the others while still joint-plan feasibility is guaranteed. Previous work on plan decoupling concentrated on a coordination framework where the only dependencies between tasks are precedence constraints. In this paper, an extension of the framework is discussed where not only precedence constraints, but also synchronisation constraints can be used in order to express qualitative temporal constraints between tasks. It is shown that adding synchronisation constraints does not add any complexity to the plan-decoupling problem, and that a previously-developed approximation algorithm for plan decoupling can be extended to cope with synchronisation constraints as well.

Formal verification of wastewater treatment processes using events detected from continuous signals by means of artificial neural networks. Case study: SBR plant

This paper proposes a modular architecture for the analysis and the validation of wastewater treatment processes. An algorithm using neural networks is used to extract the relevant qualitative patterns, such as “apexes”, “knees” and “steps”, from the signals acquired in the reaction tanks. These patterns, which show changes in the signals trend, are mapped to events in the process and logged using an appropriate XML format. The logs, in turn, are considered traces of the execution of a manufacturing process and validated using tools commonly applied for the Verification of Business Processes. The system has been applied to the data collected from a Sequencing Batch Reactor (SBR) for municipal wastewater treatment, equipped with probes for the on-line acquisition of signals such as pH, oxidation--reduction potential (ORP) and dissolved oxygen (DO). A SBR has turned out to be a suitable case study since the commonly acknowledged criteria for monitoring the biological processes (nitrification and denitrification) can be expressed in the form or qualitative constraints, which are easily translated into formal rules. The process logs, hence, are matched against these rules, which act as filters and quality classifiers.

Learning Bayesian network parameters under incomplete data with domain knowledge

Bayesian networks (BNs) have gained increasing attention in recent years. One key issue in Bayesian networks is parameter learning. When training data is incomplete or sparse or when multiple hidden nodes exist, learning parameters in Bayesian networks becomes extremely difficult. Under these circumstances, the learning algorithms are required to operate in a high-dimensional search space and they could easily get trapped among copious local maxima. This paper presents a learning algorithm to incorporate domain knowledge into the learning to regularize the otherwise ill-posed problem, to limit the search space, and to avoid local optima. Unlike the conventional approaches that typically exploit the quantitative domain knowledge such as prior probability distribution, our method systematically incorporates qualitative constraints on some of the parameters into the learning process. Specifically, the problem is formulated as a constrained optimization problem, where an objective function is defined as a combination of the likelihood function and penalty functions constructed from the qualitative domain knowledge. Then, a gradient-descent procedure is systematically integrated with the E-step and M-step of the EM algorithm, to estimate the parameters iteratively until it converges. The experiments with both synthetic data and real data for facial action recognition show our algorithm improves the accuracy of the learned BN parameters significantly over the conventional EM algorithm.

Backcasting performance of the emerging renewable energy sector in China

PurposeThis paper aims to explore the use of backcasting approach in dealing with the indeterminate future performance of the emerging renewable energy sector, based on the case of China. Through backcasting analysis of the emerging sector under uncertainties and concerns of the stakeholders, it acknowledges a range of hurdles ingrained in the current energy system and issues related to technology management within the Chinese context.Design/methodology/approachThis is a conceptual paper with integration of relevant contemporary knowledge.FindingsThis study points out the challenges in plausible developments for the renewable energy sector harnessing such approach. Particularly, the pursuit of end‐points under a multitude of optima can be associated with goal‐seeking embedded in backcasting for sustainable developments, whereas economic and technological constraints inherent to the processes of technological innovation in the cultural environment are recognized. Formulating an integrative framework, this study suggests the relevance of the approach of backcasting to augment dynamic policy analysis and planning for plausible developments in time, and to consequently optimize resource allocation in the renewal of necessary technological infrastructure.Originality/valueThis paper structurally reveals critical issues in the development of an emerging technology sector of growing importance and the pertinent implications to policy making in China for sustainable development given the underlying quantitative and qualitative constraints.

Using qualitative spatial reasoning in the conceptual design stage of a mechanical system

A qualitative approach is introduced to reason the spatial configuration of mechanisms in the conceptual design phase. Using the qualitative vector of fuzzy information, position vector, and qualitative constraint matrix, a reasoning structure for the spatial configuration of the mechanism is presented based on qualitative sign algebra. Assuming that complex mechanisms are composed of combinations of basic components, the reasoning structure is applied to reason about the spatial configuration of a complex mechanism. An analysis of the kinetics of the system is performed by reasoning with the vectors of basic components. An analysis of dynamic performance is presented.

Quantitative Inference by Qualitative Semantic Knowledge Mining with Bayesian Model Averaging

In this paper, we consider the problem of performing quantitative Bayesian inference and model averaging based on a set of qualitative statements about relationships. Statements are transformed into parameter constraints which are imposed onto a set of Bayesian networks. Recurrent relationship structures are resolved by unfolding in time to Dynamic Bayesian networks. The approach enables probabilistic inference by model averaging, i.e. it allows to predict probabilistic quantities from a set of qualitative constraints without probability assignment on the model parameters. Model averaging is performed by Monte Carlo integration techniques. The method is applied to a problem in a molecular medical context: We show how the rate of breast cancer metastasis formation can be predicted based solely on a set of qualitative biological statements about the involvement of proteins in metastatic processes.

Verification of qualitative [formula omitted] constraints

We introduce an LTL-like logic with atomic formulae built over a constraint language interpreting variables in Z . The constraint language includes periodicity constraints and comparison constraints of the form x = y and x < y ; it is closed under Boolean operations and admits a restricted form of existential quantification. Such constraints are used, for instance, in calendar formalisms or abstractions of counter automata by using congruences modulo some power of two. Indeed, various programming languages perform arithmetic operators modulo some integer. We show that the satisfiability and model-checking problems (with respect to an appropriate class of constraint automata) for this logic are decidable in polynomial space improving significantly known results about its strict fragments. This is the largest set of qualitative constraints over Z known so far, shown to admit a decidable LTL extension.

Tractable Fragments of Fuzzy Qualitative Algebra

In this paper we study the computational complexity of Fuzzy Qualitative Temporal Algebra (QA fuz ), a framework that combines qualitative temporal constraints between points and intervals, and allows modelling vagueness and uncertainty. Its tractable fragments can be identified by generalizing the results obtained for crisp Constraint Satisfaction Problems (CSPs) to fuzzy CSPs (FCSPs); to do this, we apply a general methodology based on the notion of α-cut. In particular, the results concerning the tractability of Qualitative Algebra QA, obtained in a recent study by different authors, can be extended to identify the tractable algebras of the fuzzy Qualitative Algebra QA fuz in such a way that the obtained set is maximal, namely any maximal tractable fuzzy algebra belongs to this set.

Trade-off between stability and bias in a history match problem using smoothness constraint

The history matching (HM) hydrocarbon reservoir inverse problem is ill-posed because its solution may be non-unique or unstable. Variability of acceptable solutions around local minima may be very large. Nonetheless, it is a common practice in HM to find a unique reservoir model, although there is no guarantee that this model represents the geology. Qualitative geological information is generally not considered because of difficulties to express them mathematically. Here, we incorporate ‘smoothness’ in the spatial variation of physical properties as an example of a geological qualitative constraint that can be mathematically incorporated in an objective function also honouring the data. The constraint is valid if lateral continuity exists (e.g. fluvial-deltaic siliciclastic reservoirs). We mean smoothness by conditioning the permeability and/or porosity difference between adjacent grid blocks to be small. We use a synthetic 2D water–oil model and a solution search technique allowing characterising both the optimum solution and its variability. The smoothness constraint reduces the variance of the estimates by introducing bias in the solutions still preserving good match. The key point to achieve an optimum trade-off between stability and data match is the tuning of the parameter, controlling the relative importance of the constraint in the objective function. The smoothness constraint cannot be applied to all reservoirs. However, we corroborate the idea that a ‘tool box’ of HM can be designed; each ‘tool’ can incorporate a different constraint. Therefore, the interpreter can judiciously choose a specific tool from this tool box according to the adherence of its constraint to the particular reservoir being studied.

Causal temporal constraint networks for representing temporal knowledge

In this work we describe causal temporal constraint networks (CTCN) as a new computable model for representing temporal information and efficiently handling causality. The proposed model enables qualitative and quantitative temporal constraints to be established, introduces the representation of causal constraints, and suggests mechanisms for representing inexact temporal knowledge. The temporal handling of information is achieved by structuring the information in different interpretation contexts, linked to each other through an inference mechanism which obtains interpretations that are consistent with the original temporal information. In carrying out inferences, we take into account the temporal relationships between events, the possible inexactitude associated with the events, and the atemporal or static information which affects the interpretation pattern being considered. The proposed schema is illustrated with an application developed using the CommonKADS methodology.

Constraint-Based Modeling and Kinetic Analysis of the Smad Dependent TGF-β Signaling Pathway

BackgroundInvestigation of dynamics and regulation of the TGF-β signaling pathway is central to the understanding of complex cellular processes such as growth, apoptosis, and differentiation. In this study, we aim at using systems biology approach to provide dynamic analysis on this pathway.Methodology/Principal FindingsWe proposed a constraint-based modeling method to build a comprehensive mathematical model for the Smad dependent TGF-β signaling pathway by fitting the experimental data and incorporating the qualitative constraints from the experimental analysis. The performance of the model generated by constraint-based modeling method is significantly improved compared to the model obtained by only fitting the quantitative data. The model agrees well with the experimental analysis of TGF-β pathway, such as the time course of nuclear phosphorylated Smad, the subcellular location of Smad and signal response of Smad phosphorylation to different doses of TGF-β.Conclusions/SignificanceThe simulation results indicate that the signal response to TGF-β is regulated by the balance between clathrin dependent endocytosis and non-clathrin mediated endocytosis. This model is useful to be built upon as new precise experimental data are emerging. The constraint-based modeling method can also be applied to quantitative modeling of other signaling pathways.

Interactive geophysical inversion using qualitative geological constraints

Numerical inversion of geophysical data does not normally require user interaction apart from the selection of initial inversion parameters. However, such an inversion often returns a single solution based upon default parameters. While this solution will be geophysically correct, assuming convergence of the algorithm, it may not be the most geologically reasonable answer. It is necessary to incorporate human interaction in selecting inversion solutions, this being the most efficient method for adding qualitative geological constraints. An automatic system provides a user-directed search of the space of geophysical solutions. Rankings assigned to numerical inversion results guide a genetic algorithm in advancing towards a conceptual target. Our example uses resistivity and chargeability data from a pole-dipole induced polarisation survey collected during a mineral exploration program. We invert for specific geological features: a defined, conductive top layer, sharp geological boundaries in the resistivity, and greatest depth of resolution of the inversion algorithm. The interactive system is an organised way to investigate the solution space for valid inversion results that emphasise these geological possibilities.

Constraints on multiple initial embedding of clauses

The received view is that there are no constraints on clausal embedding complexity in sentences. This hypothesis will be challenged here on empirical grounds from the viewpoint of multiple initial embedding of clauses. The data come from the British National Corpus, Brown, LOB, and philological scholarship. The results extend to several other ‘Standard Average European’ (SAE) languages like Finnish, German, Latin, and Swedish. There is a precise quantitative constraint on the degree of initial clausal embedding, and that limit is two. In double initial embeddings, a qualitative constraint prescribes that typically the highest embedded clause is an if-clause. The lower embedded clause should be the sentential subject of the if-clause. Here is a real example of a maximally complex, prototypical, initial clausal embedding in mainstream SAE: [Main [Init–1 If [Init–2 what is tantamount to dictatorship …] continues in a union] it can …] (LOB). Multiple initial self-embeddings are prohibited.

Use of qualitative constraints in modelling of the Lake Glumso

This paper describes modelling of time behaviour of phytoplankton and zooplankton in the Danish lake Glumso with a recently developed approach to machine learning in numerical domains, called Q2 learning. An essential part of this approach is qualitative constraints which were either handcrafted using knowledge from the Lotka-Volterra predator-prey model or induced directly from the collected data with a program called QUIN. The induced models were evaluated by a domain expert. We performed a comparison between numerical results of the Q2 learning approach and standard machine learning algorithms. The results suggest that use of qualitative constraints leads to more accurate quantitative predictions.

Formula omitted] Prediction of ozone concentrations

We describe a case study in which we applied Q 2 learning ( qualitatively faithful quantitative learning) to the analysis and prediction of ozone concentrations in the cities of Ljubljana and Nova Gorica, Slovenia. We used program QUIN to induce a qualitative model from numerical data that include the measurements of several meteorological and chemical variables. The resulting qualitative model consists of tree-structured monotonic qualitative constraints. We show how this model for Nova Gorica enables a nice interpretation of complex meteorological and chemical processes that affect the level of ozone concentration. In addition to inducing a qualitative model from data, we extended the qualitative model to also enable numerical prediction for both cities. In this case, we used in addition to measured data also data from the European meteorological prognostic model ALADIN which itself does not model pollutants. The results suggest that the qualitatively constrained numerical model tends to improve numerical prediction in comparison with some standard numerical learning methods.

The origins of glacially related soft-sediment deformation structures in Upper Ordovician glaciogenic rocks: implication for ice-sheet dynamics

In the upper Ordovician glaciogenic rocks of West Gondwana, many soft-sediment deformation structures are attributed to glaciotectonic processes. However, little is known about the subglacial and ice-marginal processes associated with this ice sheet. A systematic description and interpretation of these soft-sediment deformation structures is supplemented by a comparison with Late Quaternary and Modern glacial settings to provide qualitative constraints on the character of ancient subglacial thermal regimes. Soft-sediment striations were formed by discontinuous, shear-induced deformation beneath an ice sheet that oscillated between warm and cold-based thermal regimes. Beneath warm-based ice, coupling across the ice–bed interface was poor with sliding and loading occurring at this surface. Loading of soft, waterlogged sediments allowed the development of fluted surfaces, which are also associated with the formation of truncated folded zones. During times of cold thermal regimes, coupling across the ice–bed interface was enhanced by freezing, which caused changes in the position of the freezing front. These discontinuous changes, coupled with sustained subglacial shearing, resulted in the development of stacked, soft-sediment striated surfaces. In areas where the subglacial substrate had an irregular topography, complex shear zones formed that comprised stacked duplex systems interlayered with highly strained units preserving pervasive, folded lineations. At the ice front, large glaciotectonic deformation structures (comparable with push moraines) developed during sustained periods of (re)advance that created composite thrust and fold systems. Deformation was transmitted for up to 1 km from the ice front and reduced in intensity with distance. Other deformation structures include dome-like folds that consist of central anticlines (30 to 70 m in width) with tight rim synclines that formed by the load-induced, diapiric intrusion of muds. An ice-sheet surging mechanism is inferred for the development of composite thrust and fold systems and the coeval sediment diapirs. These surges potentially caused the widespread reorganisation of a glacial load resulting in the pro- to subglacial extrusion of soft sediments.

Integrating quantitative and qualitative fuzzy temporal constraints

In this work we address the problem of representing and reasoning with temporal knowledge in a very general and flexible manner. To this aim we propose a model of integration of quantitative and qu...

Improving bandwidth selection methods by adding qualitative constraints

In the context of nonparametric density estimation, we consider the combination of automatic bandwidth selection rules with qualitative constraints on the search space of bandwidths derived from bounds on the number of modes. These constraints can be easily combined with an upper bound based on the concept of oversmoothing introduced by Terrell and Scott (1985). Rather obviously, if a correct upper bound on the number of modes is known, our proposed approach helps to ensure an adequate representation of known qualitative features by the estimate. More surprisingly, even loose upper bounds on the number of modes are able to improve the MISE behavior of least squares cross-validation, by reducing the known tendency of under-smoothing of this bandwidth selector.

Qualitatively faithful quantitative prediction

We describe an approach to machine learning from numerical data that combines both qualitative and numerical learning. This approach is carried out in two stages: (1) induction of a qualitative model from numerical examples of the behaviour of a physical system, and (2) induction of a numerical regression function that both respects the qualitative constraints and fits the training data numerically. We call this approach Q 2 learning, which stands for Qualitatively faithful Quantitative learning. Induced numerical models are “qualitatively faithful” in the sense that they respect qualitative trends in the learning data. Advantages of Q 2 learning are that the induced qualitative model enables a (possibly causal) explanation of relations among the variables in the modelled system, and that numerical predictions are guaranteed to be qualitatively consistent with the qualitative model which alleviates the interpretation of the predictions. Moreover, as we show experimentally the qualitative model's guidance of the quantitative modelling process leads to predictions that may be considerably more accurate than those obtained by state-of-the-art numerical learning methods. The experiments include an application of Q 2 learning to the identification of a car wheel suspension system—a complex, industrially relevant mechanical system.

Reasoning with Numeric and Symbolic Time Information

Representing and reasoning about time is fundamental in many applications of Artificial Intelligence as well as of other disciplines in computer science, such as scheduling, planning, computational linguistics, database design and molecular biology. The development of a domain-independent temporal reasoning system is then practically important. An important issue when designing such systems is the efficient handling of qualitative and metric time information. We have developed a temporal model, TemPro, based on the Allen interval algebra, to express and manage such information in terms of qualitative and quantitative temporal constraints. TemPro translates an application involving temporal information into a Constraint Satisfaction Problem (CSP). Constraint satisfaction techniques are then used to manage the different time information by solving the CSP. In order for the system to deal with real time applications or those applications where it is impossible or impractical to solve these problems completely, we have studied different methods capable of trading search time for solution quality when solving the temporal CSP. These methods are exact and approximation algorithms based respectively on constraint satisfaction techniques and local search. Experimental tests were performed on randomly generated temporal constraint problems as well as on scheduling problems in order to compare and evaluate the performance of the different methods we propose. The results demonstrate the efficiency of the MCRW approximation method to deal with under constrained and middle constrained problems while Tabu Search and SDRW are the methods of choice for over constrained problems.