English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In cooperative multi-agent reinforcement learning (MARL) settings, thecentralized training with decentralized execution (CTDE) becomes customaryrecently due to the physical demand. However, the most dilemma is theinconsistency of jointly-trained policies and individually-optimized actions.In this work, we propose a novel value-based multi-objective learning approach,named Tchebycheff value decomposition optimization (TVDO), to overcome theabove dilemma. In particular, a nonlinear Tchebycheff aggregation method isdesigned to transform the MARL task into multi-objective optimal counterpart bytightly constraining the upper bound of individual action-value bias. Wetheoretically prove that TVDO well satisfies the necessary and sufficientcondition of individual global max (IGM) with no extra limitations, whichexactly guarantees the consistency between the global and individual optimalaction-value function. Empirically, in the climb and penalty game, we verifythat TVDO represents precisely from global to individual value factorizationwith a guarantee of the policy consistency. Furthermore, we also evaluate TVDOin the challenging scenarios of StarCraft II micromanagement tasks, andextensive experiments demonstrate that TVDO achieves more competitiveperformances than several state-of-the-art MARL methods.

TVDO: Tchebycheff Value-Decomposition Optimization for Multi-Agent Reinforcement Learning