Bridge Connecting Multitask and Dynamic: A New Approach for Multitask Optimization via Dynamic Optimization

Multitask optimization problem (MTOP) and dynamic optimization problem (DOP) widely exist in real-world applications and have been extensively studied within the field of evolutionary computation (EC). Although MTOPs and DOPs are typically studied independently, they share a fundamental characteristic that both require EC algorithms to solve multiple optimization problems within the limited fitness evaluations. This valuable relationship has not yet received sufficient attention in the EC community. In this article, we make the first attempt to bridge MTOPs to DOPs by proposing a transformation framework that reformulates an MTOP as a DOP. Through this bridge, multitask optimization is achieved by sequentially optimizing tasks in dynamic environments. The main contributions of this work are as follows. First, we propose the first transformation framework that connects the MTOP to the DOP, providing an innovative and general paradigm for multitask optimization. Second, we demonstrate the feasibility of this bridge through a concrete transformation method, which enables existing DOP algorithms to be readily applied to MTOPs. Third, experimental studies validate the effectiveness of the proposed framework in solving MTOPs via the DOP algorithm. Finally, this new perspective is expected to inspire more diverse and in-depth research into both MTOP and DOP.