Abstract: The resource placement problem is often formulated as a classical optimization problem: place N proxies within the Internet so as to minimize the average user response time for retrieving Web objects.Approaches to this problem in the literature include graph theoretic approaches, combinatorial approaches, dynamic programming approaches, and vector quantization approaches.In this paper service level agreement is induced to compose resource consumption and user’s quality requirement to one benefit function, and change the dynamic deployment problem to an optimize problem of benefit object function’s value.To solve the optimized problem, an biology imitation algorithm is presented by constructing mapping from biology action follow with the food to the problem.By imitating the biology action including competition, moving and dying, the algorithm realizes the dynamic resource deployment strategy in grid environment.