Neighborhood Interference Research Articles

Exploiting neighborhood interference with low-order interactions under unit randomized design

Abstract Network interference, where the outcome of an individual is affected by the treatment assignment of those in their social network, is pervasive in real-world settings. However, it poses a challenge to estimating causal effects. We consider the task of estimating the total treatment effect (TTE), or the difference between the average outcomes of the population when everyone is treated versus when no one is, under network interference. Under a Bernoulli randomized design, we provide an unbiased estimator for the TTE when network interference effects are constrained to low-order interactions among neighbors of an individual. We make no assumptions on the graph other than bounded degree, allowing for well-connected networks that may not be easily clustered. We derive a bound on the variance of our estimator and show in simulated experiments that it performs well compared with standard estimators for the TTE. We also derive a minimax lower bound on the mean squared error of our estimator, which suggests that the difficulty of estimation can be characterized by the degree of interactions in the potential outcomes model. We also prove that our estimator is asymptotically normal under boundedness conditions on the network degree and potential outcomes model. Central to our contribution is a new framework for balancing model flexibility and statistical complexity as captured by this low-order interactions structure.

Cooperative Transmission Mechanism Based on Revenue Learning for Vehicular Networks

With the rapid development of science and technology and the improvement of people’s living standards, vehicles have gradually become the main means of travel. The increase in vehicles has also brought about an increasing incidence of car accidents. In order to reduce traffic accidents, many researchers have proposed the use of vehicular networks to quickly transmit information. As long as these vehicles can receive information from other vehicles or buildings nearby in a timely manner, they can avoid accidents. In vehicular networks, the traditional double connection technique, through interference coordination scheduling strategy based on graph theory, can ensure the fairness of vehicles and obtain suitable neighborhood interference resistance with limited computing resources. However, when a base station transmits data to the vehicular user, the nearby base station and the vehicular network user may be in a state of suspended communication. Thus, the resource utilization of the above double connection vehicular network is not sufficient, resulting in a waste of resources. To solve this issue, this paper presents a study based on earnings learning with a vehicular network multi-point collaborative transmission mechanism, in which the vehicular network users communicate with the surrounding collaborative transmission. We use the Q-learning algorithm in the reinforcement learning process to enable vehicular network users to learn from each other and make cooperative decisions in different environments. In reinforcement learning, the agent makes a decision and changes the state of the environment. Then, the environment feeds back the benefit to the agent through the related algorithm so that the agent gradually learns the optimal decision. Simulation results demonstrate the superiority of our proposed approach with the revenue machine learning model compared with the benchmark schemes.

Numerical Investigation of Wind Pressure Coefficients on Different Scallop Dome Configurations

The effects of the action of wind on scallop domes were numerically investigated using the software Ansys, as well the interference of the neighborhood on the external pressure coefficient and the streamlines between geometrically identical domes. The influence of the proportion on the neighborhood interference in scallop domes and the variations in the dimensions of the structures on the pressure coefficients and streamlines were also investigated. Five simulations were analyzed involving six-grooved domes and geometry height variations for validation. The numerically obtained coefficients were compared with values in the literature. Other applications investigated the influence of grooves on the external pressure coefficient and the effect of wind on the grooved domes. Another application analyzed the interference of the neighborhood on the external pressure coefficients and streamlines between three geometrically identical domes and, finally, the influence of the proportion in the study of the interference of the neighborhood. Here, the variations in the dimensions of the structure affected the pressure coefficients and the streamlines were analyzed. It was possible to verify the versatility and efficiency of the computational method used in the analysis of the action of wind.

Causal Inference Under Approximate Neighborhood Interference

This paper studies causal inference in randomized experiments under network interference. Commonly used models of interference posit that treatments assigned to alters beyond a certain network distance from the ego have no effect on the ego's response. However, this assumption is violated in common models of social interactions. We propose a substantially weaker model of “approximate neighborhood interference” (ANI) under which treatments assigned to alters further from the ego have a smaller, but potentially nonzero, effect on the ego's response. We formally verify that ANI holds for well‐known models of social interactions. Under ANI, restrictions on the network topology, and asymptotics under which the network size increases, we prove that standard inverse‐probability weighting estimators consistently estimate useful exposure effects and are approximately normal. For inference, we consider a network HAC variance estimator. Under a finite population model, we show that the estimator is biased but that the bias can be interpreted as the variance of unit‐level exposure effects. This generalizes Neyman's well‐known result on conservative variance estimation to settings with interference.

珍稀濒危植物长白松(Pinus sylvestris var. sylvestriformis)天然种群生存压力

为了揭示珍稀濒危植物长白松（Pinus sylvestris var. sylvestriformis）天然种群生存压力状况，在全面调查长白山国家级自然保护区长白松分布的基础上，基于邻体干扰模型，引入树高、冠幅、方位等因子，提出3种生存压力指数：个体生存压力指数、种群生存压力指数和群落生存压力指数，分析天然长白松所处6种群落类型中的生存压力。结果表明：长白松承受群落生存压力（PI）从大到小依次为：白桦-臭冷杉群落（PI=21.532）、红松-长白松群落（PI=14.185）、白桦群落（PI=13.262）、臭冷杉-长白松群落（PI=8.752）、长白落叶松-鱼鳞云杉群落（PI=7.780）和蒙古栎群落（PI=5.440）。多重比较单向方差分析表明，6种群落类型中长白松生存压力总体上差异明显，白桦-臭冷杉群落中长白松生存压力最大，显著高于其他5种群落；竞争树种主要为长白落叶松、红松、长白松、山杨和白桦，这5个树种生存压力大小占群落生存压力的87%；红松-长白松群落和白桦群落中长白松生存压力无明显差异，但显著高于臭冷杉-长白松群落、长白落叶松-鱼鳞云杉群落和蒙古栎群落；臭冷杉-长白松群落、长白落叶松-鱼鳞云杉群落和蒙古栎群落中长白松生存压力相对较小，彼此无明显差异。长白松生存压力与其所处植物群落演替阶段及其龄级结构有关。目前，保护区采取严格保护和管理方式不完全有利于长白松种群的稳定发展。根据长白松种群所处的植物群落生境特点、种群生存压力状况并结合种群年龄结构特征，针对不同群落类型提出相应抚育措施建议以期为长白松天然种群的保护提供参考。;In order to reveal the rare and endangered plant Pinus sylvestris var. Sylvestriformis natural population survival pressure status, we made a comprehensive survey of distribution of Pinus sylvestris var. sylvestriformis population by setting up quadrats according to the altitude gradient in Changbai Mountain National Nature Reserve. We proposed three kinds of survival pressure index including individual survival pressure index, population survival pressure index and community survival pressure index to analyze the survival pressure of Pinus sylvestris var. Sylvestriformis in six types of plant communities. Ecological factors such as tree height, canopy width and tree azimuth were introduced into the pressure survival index based on neighborhood interference model. Results showed that the survival pressure of Pinus sylvestris var. Sylvestriformis presented obvious difference in six types of plant communities. The survival pressure from more to less in proper order were Betula platyphylla-Abies nephrolepis community, Pinus koraiensis-Pinus sylvestris var. sylvestriformis community, Betula platyphylla community, Abies nephrolepis-Pinus sylvestris var. Sylvestriformis community, Larix olgensis-Picea jezoensis community, and Quercus mongolica community. Multiple comparative one-way analysis of variance showed that there were significant differences in the survival pressure among the six communities. The survival pressure in Betula platyphylla-Abies nephrolepis community was the highest, which was significantly higher than that in other five communities. The main competitive species were Larix olgensis, Pinus koraiensis, Pinus sylvestris var. sylvestriformis, Populus davidiana, and Betula platyphylla, which accounted for 87% of the total community survival pressure of Pinus sylvestris var. Sylvestriformis. There was no significant difference in the survival pressure of Pinus koraiensis-Pinus sylvestris var. sylvestriformis community and Betula platyphylla community, but the survival pressure of Pinus koraiensis-Pinus sylvestris var. sylvestriformis community was significantly higher than that of Abies nephrolepis-Pinus sylvestris var. Sylvestriformis community, Larix olgensis-Picea jezoensis community, and Quercus mongolica community. Among the latter three communities, there was no significant difference in the survival pressure. Pinus sylvestris var. Sylvestriformis was a strongly heliophilous species, the survival pressure of Pinus sylvestris var. Sylvestriformis was related to the plant communities succession stage and age structure of Pinus sylvestris var. Sylvestriformis population. The strict protection and management in the Changbai Mountain Nature Reserve was not completely conducive to the stable development of the population of Pinus sylvestris var. Sylvestriformis. Finally, according to the characteristics of plant community habitat, population survival pressure and population age structure characteristics, we put forward the corresponding management measures for six types of plant communities to provide reference for the protection of the population of Pinus sylvestris var. Sylvestriformis.

A quality of system routing based on neighborhood interference detection and mitigation in wireless sensor network

The current quality of system routing available in wireless sensor network mainly provides assurance for traffic flow, with little consideration on full use of network resource, like bandwidth. Aiming at solving these problems, this article proposes a quality of system routing method based on neighborhood interference detection and mitigation. First, the algorithm uses a sink node as root node to establish a minimum spanning tree for the network to obtain the depth of each node. Then, each node dynamically updates its interference degree by contiguous iterations during neighboring nodes communication. Finally, the interference degree, packet level, residual energy, and node depth are combined to establish a multipath route to enhance the network bandwidth and thus improve the quality of system. Experimental results show that the algorithm can not only effectively balance the network load but also reduce the network packet delay and improve the network throughput. It also has a good network lifetime; thus, it can effectively improve the quality of system of wireless sensor network.

Causal Inference Under Approximate Neighborhood Interference

Causal Inference Under Approximate Neighborhood Interference

Interference-aware multipath routing in wireless mesh network

The inherent mesh infrastructure of IEEE 802.11-based wireless mesh networks provides added support to construct multiple robust paths. However, based on geometric locations of wireless nodes, neighborhood interference and channel contention impose challenges on multipath routing schemes. A large body of research has been carried out to maximize aggregate end-to-end throughput using multipath routing; however, interference has not been accurately modeled in majority of the work. Based on the relative location of transmitters and receivers, information asymmetric non-coordinated interference introduces bottleneck links and significantly reduces the aggregate throughput. The impact of this interference is even observed on links several hops away. In this paper, multipath routing is integrated with topology control to manage such multilevel asymmetric interference. An optimization model has been presented with an objective to achieve optimized end-to-end throughput using multiple available paths, considering coordinated and asymmetric non-coordinated interference. The goal of this research is to develop a multipath routing strategy which can achieve better end-to-end throughput by purging badly affected asymmetric non-coordinated interfering links during path construction procedure. The proposed model and routing strategy have been tested through extensive simulations. The results clearly exhibit the efficacy of the proposed approach, which achieves better aggregate end-to-end throughput compared to existing multipath routing schemes.

Response of gas exchange to neighborhood interference in leaves of teak ( Tectona grandis L. f.) in a tropical plantation forest

The growth performance of individual plants in a population can be affected by the plant-plant interaction, which has been well recognized and is called neighborhood interference. Though mechanisms are still unclear, the variations in gas exchange parameters in relation to the neighborhood interference between individual plants are crucial for evaluating the effects of plant-plant interaction. CO 2 assimilation in leaves of teak under natural conditions of dry seasons and wet seasons as well as its response to variations in light flux density and CO 2 concentration in different neighborhood interferences were simultaneously measured with Li-6400 portable photosynthesis system in a 21-year-old tropical plantation forest in Jianfengling, Hainan Island, China. This article dealt with the change rule of neighborhood interference on tree characteristics of gas exchange and its dynamic response to light en-vironments of individual plants. Diurnal courses of photosynthesis of individual leaves were not affected by neighborhood interfer-ence, but net photosynthetic rates showed a negative relationship with the intensity of neighborhood interference. The ratio of daily average P n in weak, moderate, strong and heavy neighborhood interferences was 2.5:2.3:1.7:1.0, and the daily maximum P n in weak interference was 2.8 times that in heavy interference. Leaf transpiration and stomatal conductance decreased with the increasing of interference intensity. Characteristics of photosynthetic light response and CO 2 response changed with the neighborhood interference, and values of leaf gas exchange parameters including A sat, Q sat, α A, CE, V cmax and J max in weak interference were enhanced by 2.7, 1.3, 1.4, 2.7, 1.9 and 2.8 times, respectively, than those in heavy interference. But changes in those parameters partly depended on light environment and CO 2 concentration, and the influence of changes in light environment on weak interference individuals was sig-nificantly stronger than that on heavy interference. Beyond the growth CO 2 concentration, the influence resulting from changes in CO 2 concentration on heavy interference individuals was obviously stronger than that on weak interference. Neighborhood interference can be described as a major means of intraspecific competition of population in a plantation forest with uniform forest structure and consistent management. Carbon assimilation can be affected by the neighborhood interference, and result in divergence in growth performance. Indices of neighborhood interference can be used to evaluate the intraspecific competition, and based on this point, we can make maximum use of resources after stand structure has been well adjusted.

Application of the modified neighborhood interference model to the cultivation of forest and herbaceous ecosystem

This paper is an extension of the application of neighborhood interference model—being modified—to the study of the interference between forest plants and herbaceous plants. The exploration of the neighborhood interference between triploid Populus tomentosa and Lolium multiflorum shows that the interference of Lolium multiflorum on triploid populus tomentosa is positive, with the primary interference factor living weight, and secondary factor cover degree; the growth of an individual tree is affected by the number, height, and coverage of the Lolium multiflorum; the more dense the coverage of the Lolium multiflorum in a certain area, the taller, faster and more trees will grow as a result of a greater positive neighborhood interference. The modified neighborhood interference model well serves as an effective approach to extend the application of neighborhood interference researches.

Growth of tropical savanna grass plants in competition: A shoot population model

The Venezuelan savannas, and in general Neotropical savannas, are composed of herbaceous co-dominant species and less abundant trees and shrubs, conditioned in great measure by the occurrence of annual fire events. We studied the growth dynamics of three grass species representative of the major phenological and architectural types: a precocious, bunch grass, Elyonurus adustus, an early, medium tall grass Leptocoryphium lanatum and a late, tall species Andropogon semiberbis. The experimental design included individuals of all these species growing alone, interacting in species pairs and all three species in competitive interaction. Close inspection of the data showed that the species differ widely in the spatial and temporal patterns of occupation of aboveground space, showing important levels of neighborhood interference among them. Shoot emergence during the time period between fire events (usually 1 year) is not a continuous process but occurs as a set of discrete cohorts. We model plant growth as a population of shoots governed by ordinary differential equations that accounts for the existence of different number of cohorts as a consequence of shoot competition. A comparison of simulated patterns to those observed in the field was assessed by the root mean square (RMS) error between the transient behaviors of both patterns. L. lanatum resulted both the most impaired and the least impairing of the three species and the opposite is valid for A. semiberbis. For the former species, the competition effect when growing together with the other two is stronger that when growing in pairs and close to an additive process. For E. adustus and A. semiberbis the addition of the third species does not substantially modified the effect obtained when competing in pairs. This model can be applied to the analysis of the herbaceous plant growth process on savannas subject to fire. With modifications, the model can be applied to other grass plants and other grassland ecosystems. We suggest that a possible mechanism involved in this process is the growth inhibition of the basal shoots by the older shoots of the same plant.

Neighborhood predictors of plant performance

We developed models of inter-individual interference to predict the fecundity of individuals in populations of the annual plant species, Arabidopsis thaliana. An individual plant is modeled as having a neighborhood which is a circular area of fixed radius with the plant at its center. Other plants which share the circle with the focal plant are termed neighbors of the focal plant. We developed an index of neighborhood interference which is the independent variable in a non-linear regression model that predicts individual plant fecundity. We present methods of exploratory data analysis that are useful in determining a best neighborhood radius, defined as that radius which minimizes residual sum of squares, and in deciding on the functional form of the interference index. In developing the interference index for Arabidopsis, we focus on aspects of the spatial distribution of neighbors: their number, distance and angular dispersion.We found that a best (or optimal) neighborhood radius can be resolved, which provides the best predictor of plant performance. Fecundity predictors based on adult neighbors were noticeably better than those based on neighbors at the seedling stage. Rosettes of Arabidopsis may change location during development (they fall over) and the new "fallen" positions do provide some improvement in the predictor. Taking into account distance to neighbors within the neighborhood provided only negligible improvement in the model. Finally, the incorporation of angular dispersion in the crowding index produced a considerably better fit. The fecundity predictor that included number of neighbors and angular dispersion in the crowding index explained about 70% of the variation in individual seed set.