基于事件的社交网络上的双边偏好稳态规划

在基于事件的社交网络中,一个经典的问题是为用户规划其感兴趣的事件.现有的工作仅仅考虑用户的喜好,仅从用户的角度出发,为其安排尽可能感兴趣的事件来参加.然而,从事件主办者的角度出发,他们亦希望为事件安排的用户尽可能有更大的影响力,用户的可靠性尽可能高,以保障事件能够顺利开展,并取得预期的效果.本质上来说,基于事件的社交网络上的规划问题是一个双向选择的问题,而现有的所有工作均未从用户和事件的双边偏好考虑问题.因此,提出一种双边偏好稳态规划问题来解决这种双向选择问题.该问题首次提出,因此现有工作中未有相关算法可供解决该问题.对比之前只考虑用户偏好的规划,在考虑用户和事件双边偏好时,面临着问题更复杂、约束条件更多的困难.因此,提出两种基础算法和一种改进算法来高效、高质量地解决这个问题,并用大量的实验验证所提出算法的高效性和有效性.

Analyses of Valid Range for the Linear Approximation in a Single-Mode Laser

Using the linear approximation method, we calculated the steady-state mean normalized intensity fluctu-ation for a loss-noise model of a single-mode laser driven by a pump noise and a quantum noise, whose real part andimaginary part are cross-correlated. We analyzed the valid range for thelinear approximation method by studying theinfluences on the steady-state mean normalized intensity fluctuation by the cross-correlation coefficient, the intensities ofthe quantum and pump noise, the net gain, and the amplitude and frequency of the input signal, and we found that thevalid range becomes wider when the cross-correlation between the real and imaginary part of quantum noise is weaker,the noise intensities of quantum and pump are weaker, the laser system is far from the threshold and the signal hassmaller amplitude and higher frequency.