This paper examines the objective of optimally harvesting a single species in a stochastic environment. This problem has previously been analyzed in Alvarez (2000) using dynamic programming techniques and, due to the natural payoff structure of the price rate function (the price decreases as the population increases), no optimal harvesting policy exists. This paper establishes a relaxed formulation of the harvesting model in such a manner that existence of an optimal relaxed harvesting policy can not only be proven but also identified. The analysis embeds the harvesting problem in an infinite-dimensional linear program over a space of occupation measures in which the initial position enters as a parameter and then analyzes an auxiliary problem having fewer constraints. In this manner upper bounds are determined for the optimal value (with the given initial position); these bounds depend on the relation of the initial population size to a specific target size. The more interesting case occurs when the initial population exceeds this target size; a new argument is required to obtain a sharp upper bound. Though the initial population size only enters as a parameter, the value is determined in a closed-form functional expression of this parameter.
Existence of Relaxed Equilibrium for Time-Inconsistent Mean Field Games: A Set-Valued Fixed Point Approach
Zhongtai Securities Institute for Financial Studies, Shandong University(山东大学中泰金融研究院);School of Mathematics and Statistics, University of Sydney(悉尼大学数学与统计学院)
Spatial ecology, optimal control and game theoretical fishing problems
在生态系统和经济政策中最重要的是自然资源的收获问题。提出这一问题的典型情况是捕鱼战略。事实上,过度捕捞是管理生物资源方面的一个众所周知的问题,因为过于贪婪可能导致我们正在捕捞的种群的全面急剧枯竭。一个密切相关的专题是渔业政策方面的纳什均衡。也就是说,两个参与者在竞争同一个资源库,他们是否可能找到一个均衡的情况?本文的目标是提供一个详细的分析,这两个查询(\emph{i.e}最佳捕鱼策略的单人游戏模型和研究的纳什均衡的多个玩家游戏),使用一个基本的,但有指导意义的数学模型,物流扩散方程。在这个框架中,底层模型简单地读取$-μΔθ=θ(K(x)-α(x)-θ)$,其中$K$说明自然资源,$θ$说明被捕捞的种群密度,而$α=α(x)$编码单个玩家的捕捞策略,或者在处理纳什均衡时,编码两个玩家的捕捞策略的组合。本文由两个主要部分组成。第一个给出了单人游戏的优化器的非常好的特征。在两个参与者参与的情况下,我们的目标是找到一个纳什均衡。我们证明了纳什均衡的存在性在几个不同的制度\textcolor{black}{and investigate several related qualitative queries}。我们的研究是完成了各种数值模拟,说明我们的结果,并允许我们制定开放的问题和解答。
Accuracy of Discretely Sampled Stochastic Policies in Continuous-time Reinforcement Learning
Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Hong Kong(系统工程与工程管理系,香港中文大学);Department of Mathematical Sciences, Tsinghua University, China(数学科学系,清华大学);Department of Mathematics, Imperial College London, United Kingdom(数学系,伦敦帝国学院)