Internet of Things (IoT) and cloud computing are steadily growing in support of recent and projected revolutionary Internet applications. Cloud computing has the ability to meet the performance expectations of different applications. In this paper, we present the implementation of applications based on cooperative resource and energy-constrained objects with optimized performances. To dynamically incorporate objects into IoT applications' execution, task scheduling should be implemented for resource allocation in an optimized manner. We propose a task scheduling algorithm based on robust canonical particle swarm optimization (CPSO) and fully-informed particle swarm (FIPS) algorithms in order to solve the problem of resource allocation and management in both homogeneous and heterogeneous IoT cloud computing. Our objective is to satisfy the Quality of Service (QoS) in terms of throughput and delay, by performing optimal task scheduling taking into consideration different classes of data traffic. Performance evaluation of experiments show that throughput and delay can be significantly improved by dynamic dedicated servers (DDSS) and heterogeneous DDSS (h-DDSS) using FIPS optimization algorithm compared to CPSO optimization algorithm.