Uncertainty In Dynamic Environment Research Articles

Inference and Learning in Dynamic Decision Networks Using Knowledge Compilation

Decision making under uncertainty in dynamic environments is a fundamental AI problem in which agents need to determine which decisions (or actions) to make at each time step to maximise their expected utility. Dynamic decision networks (DDNs) are an extension of dynamic Bayesian networks with decisions and utilities. DDNs can be used to compactly represent Markov decision processes (MDPs). We propose a novel algorithm called mapl-cirup that leverages knowledge compilation techniques developed for (dynamic) Bayesian networks to perform inference and gradient-based learning in DDNs. Specifically, we knowledge-compile the Bellman update present in DDNs into dynamic decision circuits and evaluate them within an (algebraic) model counting framework. In contrast to other exact symbolic MDP approaches, we obtain differentiable circuits that enable gradient-based parameter learning.

Flexible structure learning under uncertainty.

Experience is known to facilitate our ability to interpret sequences of events and make predictions about the future by extracting temporal regularities in our environments. Here, we ask whether uncertainty in dynamic environments affects our ability to learn predictive structures. We exposed participants to sequences of symbols determined by first-order Markov models and asked them to indicate which symbol they expected to follow each sequence. We introduced uncertainty in this prediction task by manipulating the: (a) probability of symbol co-occurrence, (b) stimulus presentation rate. Further, we manipulated feedback, as it is known to play a key role in resolving uncertainty. Our results demonstrate that increasing the similarity in the probabilities of symbol co-occurrence impaired performance on the prediction task. In contrast, increasing uncertainty in stimulus presentation rate by introducing temporal jitter resulted in participants adopting a strategy closer to probability maximization than matching and improving in the prediction tasks. Next, we show that feedback plays a key role in learning predictive statistics. Trial-by-trial feedback yielded stronger improvement than block feedback or no feedback; that is, participants adopted a strategy closer to probability maximization and showed stronger improvement when trained with trial-by-trial feedback. Further, correlating individual strategy with learning performance showed better performance in structure learning for observers who adopted a strategy closer to maximization. Our results indicate that executive cognitive functions (i.e., selective attention) may account for this individual variability in strategy and structure learning ability. Taken together, our results provide evidence for flexible structure learning; individuals adapt their decision strategy closer to probability maximization, reducing uncertainty in temporal sequences and improving their ability to learn predictive statistics in variable environments.

Building community resilience through cross‐sector partnerships and interdisciplinary research

AbstractBuilding community resilience has become a national imperative. Substantial uncertainties in dynamic environments of emergencies and crises require real‐time information collection and dissemination based on big data analytics. These, in turn, require networked communities and cross‐sector partnerships to build lasting resilience. This viewpoint article highlights an interdisciplinary approach to building community resilience through community‐engaged research and partnerships. This perspective leverages existing community partnerships and network resources, undertakes an all‐hazard and whole‐community approach, and evaluates the use of state‐of‐the‐art information communication technologies. In doing so, it reinforces the multifaceted intergovernmental and cross‐sector networks through which resilience can be developed and sustained.

RRT*-Fuzzy Dynamic Window Approach (RRT*-FDWA) for Collision-Free Path Planning

Path planning is an important aspect and component in the research of mobile-robot-related technologies. Many path planning algorithms are only applicable to static environments, while in practical tasks, the uncertainty in dynamic environments increases the difficulty of path planning and obstacle avoidance compared with static environments. To address this problem, this paper proposes an RRT*-FDWA algorithm. RRT* first generates a global optimal path, and then, when obstacles exist nearby, an FDWA algorithm fixes the local path in real time. Compared with other path planning algorithms, RRT*-FDWA can avoid local minima, rapidly perform path replanning, generate a smooth optimal route, and improve the robot’s maneuvering amplitude. In this paper, the effectiveness of the algorithm is verified through experiments in dynamic environments.

Obstacle Avoidance With Dynamic Avoidance Risk Region for Mobile Robots in Dynamic Environments

Uncertainty in dynamic environment increases complexity and difficulty in obstacle avoidance of robots. An avoidance method with dynamic obstacle avoidance risk region in a dynamic environment is proposed in this study. First, using the state estimation of the extended kalman filter, a dynamic obstacle avoidance risk region is constructed along the direction of the obstacle movement. Second, the robot's safe obstacle avoidance operation against dynamic obstacles combined with nonlinear model predictive control is realized. Finally, a series of experiments are used to verify the effectiveness of the proposed method. The experimental results showed that the robot can easily control its movement and a flat obstacle avoidance trajectory is obtained via the proposed method.

Manage the unmanageable? The illusion of control of new venture teams coping with dynamic environmts

Behavioral decision theory and the hubris theory of entrepreneurship suggest that entrepreneurs' uncertainty in dynamic environments leads to cognitive biases. Despite the elaborated theory, empirical evidence of the dynamism - bias relationship is scarce and has produced unexpected results. Further, the role that mechanisms applied to cope with environmental dynamism play in the emergence of cognitive biases, remains unclear. This study investigates how the interplay of environmental dynamism and environmental scanning relates to the illusion of control bias of new venture teams (NVTs). The NVT illusion of control describes an NVT’s tendency to systematically overestimate the degree to which their skills and performance, instead of chance, will influence uncontrollable events. We test our hypotheses by applying structural equation modeling with a sample of 123 NVTs in Germany. Our findings reveal a lack of direct relationship between environmental dynamism and NVT illusion of control but discover an indirect relationship between the two mediated by environmental scanning. These findings suggest that NVTs regularly overestimate their ability to manage their largely unmanageable environment when applying the coping mechanism of environmental scanning.

Risk-Aware Path Planning Under Uncertainty in Dynamic Environments

This study develops a novel sampling-based path planning approach, simultaneously achieving uncertainty reduction of localization and avoidance of dynamic obstacles. The proposed path planner can generate a set of path primitives and the path selection takes into account the localization uncertainty, the collision-risk, and the cost-to-go to the target area. The weights of these quantities for selecting an optimal path are tuned adaptively by using the entropy weight method. In the process of path primitive generation, we propose an adaptive planning horizon scheme that can generate a longer path with lower localization uncertainty. Particularly, to further reduce the localization uncertainty of the path primitive, we propose a sampling strategy that is capable of biasing the sampling points to the information-rich areas. To reduce the collision-risk, we propose to calculate the probability of collision by taking the uncertainty of both the robot and the dynamic objects into consideration. The proposed approach and its key components are verified in extensive experiments in both simulation and real-world environments. The proposed method is demonstrated to be capable of efficiently guiding the robot to the designated location with lower localization uncertainty and higher success rate in obstacle avoidance.

Trajectory planning for unmanned surface vehicles operating under wave-induced motion uncertainty in dynamic environments

We present a deliberative trajectory planning method to avoid collisions with traffic vessels. It also plans traversal across wavefields generated by these vessels and minimizes the risk of failure. Our method searches over a state-space consisting of pose and time. And, it produces collision-free and minimum-risk trajectory. It uses a lookup table to account for motion uncertainty and failure risk. We also present speed-up techniques to increase performance. Our wave-aware planner produces plans that (1) have shorter execution times and safer when compared to previously developed reactive planning schemes and (2) comply with user-defined wave-traversal constraints and Collision Regulations (COLREGs)

Sustainable Scheduling of an Automatic Pallet Changer System by Multi-Objective Evolutionary Algorithm with First Piece Inspection

In this study, the machining center with the Automated Pallet Changer (APC) scheduling problem considering the disturbance of the first piece inspection is presented. The APC is frequently used in industry practice; it is useful in terms of sustainability and robustness because it increases the machine utilization rate and enhances the responsiveness to uncertainties in dynamic environments. An enhanced evolutionary algorithm for APC scheduling (APCEA) is developed by combining the multi-objective evolutionary algorithm with APC simulation. The dynamic factors in the simulation model include the pass rate of the first piece inspection (FPI) and the adjusted time when the FPI is unpassed. The proposed APCEA defines the non-robust gene based on the risk combination of the first piece inspection, and screens the non-robust gene in the genetic operation, thus improving the solution quality under the same computation times. Compared with the other three multi-objective evolutionary algorithms (MOEAs), it is demonstrated that the proposed APCEA produces the best result among the four methods. The proposed APCEA has been embedded into the manufacturing execution system (MES) and successfully applied in a manufacturing plant. The application value of the proposed method is verified by a practical example.

Reasoning about norms under uncertainty in dynamic environments

The behaviour of norm-autonomous agents is determined by their goals and the norms that are explicitly represented inside their minds. Thus, they require mechanisms for acquiring and accepting norms, determining when norms are relevant to their case, and making decisions about norm compliance. Up until now the existing proposals on norm-autonomous agents assume that agents interact within a deterministic environment that is certainly perceived. In practise, agents interact by means of sensors and actuators under uncertainty with non-deterministic and dynamic environments. Therefore, the existing proposals are unsuitable or, even, useless to be applied when agents have a physical presence in some real-world environment. In response to this problem we have developed the n-BDI architecture. In this paper, we propose a multi-context graded BDI architecture (called n-BDI) that models norm-autonomous agents able to deal with uncertainty in dynamic environments. The n-BDI architecture has been experimentally evaluated and the results are shown in this paper.

Project portfolios in dynamic environments: Organizing for uncertainty

This research investigated the following research question: How is uncertainty affecting project portfolios managed in dynamic environments? While different approaches have been developed in the context of the management of single projects these ideas have not been carried over to the management of project portfolios.The dynamic capabilities framework is used as the framework to study the management of project portfolios in dynamic environments. The research is based on four portfolios in two firms using retrospective analysis. Sufficient material was collected and analyzed to contribute in the following areas: (1) To provide a better understanding of the management of project portfolios facing uncertainty, (2) to analyze the relationships between the sources of uncertainty in dynamic environments and the organizing mechanisms put in place by organizations to minimize their impact and to capitalize on opportunities, and (3) to identify possible improvements to project portfolio models and standards.