An essential concern in the modern period is the growing energy consumption in the Internet of Things (IoT) networks, particularly when connecting resource-constrained devices. Prolonging the network's lifespan under these circumstances poses a significant challenge. Clustering is a prominent strategy to enhance the network's operational duration. However, poor cluster heads (CH) selection rapidly drains energy from network nodes. In multi-hop clustering systems, CH consumes more power by forwarding the data from other CHs and its data, creating "hot spots" in the network. We propose a novel evolutionary multi-objective clustering protocol (EMCP) for heterogeneous IoT networks to prolong the network lifetime. In the initial stage, the genetic algorithm optimizes the CH selection process by considering the residual energy, communication radius, the distance between the edge node and CH, and the distance between CHs as fitness parameters. The distance between IoT nodes and CHs determines the clustering process. The energy monitoring module also keeps track of the CHs' power consumption following each data transmission. If the CH's remaining power drops to a threshold value, the new cluster head is one of the IoT nodes close to the previous cluster head with the highest residual energy. In addition, the simulation is performed with multiple edge servers to solve the hot spot problem. The suggested protocol's performance is compared against state-of-the-art protocols for various parameters, including network stability, network longevity, the number of alive nodes, residual energy, and throughput. According to the simulation findings, the EMCP protocol achieves 15,347, 19,493, 24,187, 41,833 rounds, and 1,901,246 packets, respectively, for metrics such as network stability, half nodes dead, three-quarter node dead, last node dead, and throughput. Similarly, EMCP-MS has 17,565, 25,585, 30,410, 45,567 rounds, and 1,957,942 packets for the same metrics. Thus, the EMCP protocol has 94.55%, 80.54%, and 26.15% more life than the GAOC, DDMPEAwithANUM, and OptGACHE protocols. EMCP-MS has 75.49%, 56.55%, and 26.51% more longevity than multi-sink versions of the above-mentioned state-of-the-art protocols. [ABSTRACT FROM AUTHOR]