Incremental Information-Aware Heuristics for Task Offloading in Mobile Edge Computing

Task offloading in Mobile Edge Computing (MEC) involves a fundamental trade-off between some network parameters. Exact optimization can compute the optimal makespan for a given network snapshot, but is computationally prohibitive for dynamic environments. This work proposes and evaluates incremental heuristics that operate with different levels of information, ranging from strictly local decisions to cooperative strategies across multiple base stations in real-time execution time, and investigates how information granularity affects offloading performance under uncertainty. A structured experimental campaign was conducted, with results compared against the optimal makespan. The study demonstrates that the effectiveness of each heuristic depends on the interplay between link quality, load distribution, and available network information, suggesting opportunities for adaptive offloading mechanisms that dynamically select the most suitable strategy.