Abstract:  To solve some complicated function optimization problems, this paper proposes the ecological balance dynamics-based optimization (EBDO) algorithm based on the Lotka-Volterra ecological balance dynamics. The algorithm supposes that some types of autotrophy (such as plants), consumer (such as animals) and decomposer population exist in the ecosystem, consumer populations take autotrophy populations as food; decomposer populations take dead consumer populations as food and provide nutrition for autotrophy populations. This paper constructs the consumer-autotrophy operator, autotrophy-decomposer operator and decomposer-consumer operator by using the above relationship between ecosystem phenomena. The growth of autotrophy, consumer and decomposer populations is equivalent to alternative solutions transferring from a position to another in the search space of an optimization problem. Results show that the algorithm has the characteristics of strong search capability and global convergence, and has a high convergence speed for the complicated function optimization problems.

Key words: heuristic algorithm, swarm intelligence optimization computation, evolution computation, Lotka-Volterra ecological balance dynamics model

摘要： 为了解决复杂函数优化问题, 提出了一种Lotka-Volterra生态平衡动力学优化算法。该算法假设在某个生态系统中有自养者、消费者和分解者3个种群。自养者主要是植物；消费者主要是以自养者为食的动物；分解者主要分解消费者的死有机体, 并给自养者提供营养物质。根据上述生态系统中种群的关系构造出了消费者-自养者算子、自养者-分解者算子、分解者-消费者算子和生长算子。自养者、消费者和分解者种群的生长变化相当于搜索空间的试探解从一个位置转移到另外一个位置。该算法具有搜索能力强和全局收敛性的特点，为复杂优化问题的求解提供了一种解决方案。

关键词: 启发式算法, 群智能优化计算, 进化计算, Lotka-Volterra生态平衡动力学模型