In cooperative multi-agent reinforcement studying (MARL), as a consequence of its *on-policy* nature, coverage gradient (PG) strategies are usually believed to be much less pattern environment friendly than worth decomposition (VD) strategies, that are *off-policy*. Nevertheless, some latest empirical research reveal that with correct enter illustration and hyper-parameter tuning, multi-agent PG can obtain surprisingly sturdy efficiency in comparison with off-policy VD strategies.

**Why might PG strategies work so nicely?** On this submit, we’ll current concrete evaluation to point out that in sure eventualities, e.g., environments with a extremely multi-modal reward panorama, VD may be problematic and result in undesired outcomes. In contrast, PG strategies with particular person insurance policies can converge to an optimum coverage in these instances. As well as, PG strategies with auto-regressive (AR) insurance policies can study multi-modal insurance policies.

Determine 1: totally different coverage illustration for the 4-player permutation recreation.

## CTDE in Cooperative MARL: VD and PG strategies

Centralized coaching and decentralized execution (CTDE) is a well-liked framework in cooperative MARL. It leverages *world* data for simpler coaching whereas conserving the illustration of particular person insurance policies for testing. CTDE may be carried out by way of worth decomposition (VD) or coverage gradient (PG), main to 2 various kinds of algorithms.

VD strategies study native Q networks and a mixing perform that mixes the native Q networks to a world Q perform. The blending perform is often enforced to fulfill the Particular person-International-Max (IGM) precept, which ensures the optimum joint motion may be computed by greedily selecting the optimum motion domestically for every agent.

In contrast, PG strategies straight apply coverage gradient to study a person coverage and a centralized worth perform for every agent. The worth perform takes as its enter the worldwide state (e.g., MAPPO) or the concatenation of all of the native observations (e.g., MADDPG), for an correct world worth estimate.

## The permutation recreation: a easy counterexample the place VD fails

We begin our evaluation by contemplating a stateless cooperative recreation, particularly the permutation recreation. In an $N$-player permutation recreation, every agent can output $N$ actions ${ 1,ldots, N }$. Brokers obtain $+1$ reward if their actions are mutually totally different, i.e., the joint motion is a permutation over $1, ldots, N$; in any other case, they obtain $0$ reward. Observe that there are $N!$ symmetric optimum methods on this recreation.

Determine 2: the 4-player permutation recreation.

Determine 3: high-level instinct on why VD fails within the 2-player permutation recreation.

Allow us to concentrate on the 2-player permutation recreation now and apply VD to the sport. On this stateless setting, we use $Q_1$ and $Q_2$ to indicate the native Q-functions, and use $Q_textrm{tot}$ to indicate the worldwide Q-function. The IGM precept requires that

[argmax_{a^1,a^2}Q_textrm{tot}(a^1,a^2)={argmax_{a^1}Q_1(a^1),argmax_{a^2}Q_2(a^2)}.]

We show that VD can not signify the payoff of the 2-player permutation recreation by contradiction. If VD strategies had been in a position to signify the payoff, we might have

[Q_textrm{tot}(1, 2)=Q_textrm{tot}(2,1)=1quad text{and}quad Q_textrm{tot}(1, 1)=Q_textrm{tot}(2,2)=0.]

If both of those two brokers has totally different native Q values (e.g. $Q_1(1)> Q_1(2)$), we’ve $argmax_{a^1}Q_1(a^1)=1$. Then in line with the IGM precept, *any* optimum joint motion

[(a^{1star},a^{2star})=argmax_{a^1,a^2}Q_textrm{tot}(a^1,a^2)={argmax_{a^1}Q_1(a^1),argmax_{a^2}Q_2(a^2)}]

satisfies $a^{1star}=1$ and $a^{1star}neq 2$, so the joint motion $(a^1,a^2)=(2,1)$ is sub-optimal, i.e., $Q_textrm{tot}(2,1)<1$.

In any other case, if $Q_1(1)=Q_1(2)$ and $Q_2(1)=Q_2(2)$, then

[Q_textrm{tot}(1, 1)=Q_textrm{tot}(2,2)=Q_textrm{tot}(1, 2)=Q_textrm{tot}(2,1).]

Consequently, worth decomposition can not signify the payoff matrix of the 2-player permutation recreation.

What about PG strategies? Particular person insurance policies can certainly signify an optimum coverage for the permutation recreation. Furthermore, stochastic gradient descent can assure PG to converge to one in every of these optima beneath gentle assumptions. This implies that, though PG strategies are much less fashionable in MARL in contrast with VD strategies, they are often preferable in sure instances which might be widespread in real-world purposes, e.g., video games with a number of technique modalities.

We additionally comment that within the permutation recreation, with a view to signify an optimum joint coverage, every agent should select distinct actions. **Consequently, a profitable implementation of PG should be certain that the insurance policies are agent-specific.** This may be accomplished through the use of both particular person insurance policies with unshared parameters (known as PG-Ind in our paper), or an agent-ID conditioned coverage (PG-ID).

## PG outperforms present VD strategies on fashionable MARL testbeds

Going past the easy illustrative instance of the permutation recreation, we prolong our research to fashionable and extra reasonable MARL benchmarks. Along with StarCraft Multi-Agent Problem (SMAC), the place the effectiveness of PG and agent-conditioned coverage enter has been verified, we present new leads to Google Analysis Soccer (GRF) and multi-player Hanabi Problem.

Determine 4: (left) successful charges of PG strategies on GRF; (proper) greatest and common analysis scores on Hanabi-Full.

In GRF, PG strategies outperform the state-of-the-art VD baseline (CDS) in 5 eventualities. Apparently, we additionally discover that particular person insurance policies (PG-Ind) with out parameter sharing obtain comparable, typically even greater successful charges, in comparison with agent-specific insurance policies (PG-ID) in all 5 eventualities. We consider PG-ID within the full-scale Hanabi recreation with various numbers of gamers (2-5 gamers) and examine them to SAD, a powerful off-policy Q-learning variant in Hanabi, and Worth Decomposition Networks (VDN). As demonstrated within the above desk, PG-ID is ready to produce outcomes corresponding to or higher than the most effective and common rewards achieved by SAD and VDN with various numbers of gamers utilizing the identical variety of atmosphere steps.

## Past greater rewards: studying multi-modal conduct by way of auto-regressive coverage modeling

In addition to studying greater rewards, we additionally research how one can study multi-modal insurance policies in cooperative MARL. Letâ€™s return to the permutation recreation. Though we’ve proved that PG can successfully study an optimum coverage, the technique mode that it lastly reaches can extremely depend upon the coverage initialization. Thus, a pure query will probably be:

Can we study a single coverage that may cowl all of the optimum modes?

Within the decentralized PG formulation, the factorized illustration of a joint coverage can solely signify one explicit mode. Due to this fact, we suggest an enhanced technique to parameterize the insurance policies for stronger expressiveness â€” the auto-regressive (AR) insurance policies.

Determine 5: comparability between particular person insurance policies (PG) and auto-regressive insurance policies (AR) within the 4-player permutation recreation.

Formally, we factorize the joint coverage of $n$ brokers into the type of

[pi(mathbf{a} mid mathbf{o}) approx prod_{i=1}^n pi_{theta^{i}} left( a^{i}mid o^{i},a^{1},ldots,a^{i-1} right),]

the place the motion produced by agent $i$ relies upon by itself statement $o_i$ and all of the actions from earlier brokers $1,dots,i-1$. The auto-regressive factorization can signify *any* joint coverage in a centralized MDP. The *solely* modification to every agentâ€™s coverage is the enter dimension, which is barely enlarged by together with earlier actions; and the output dimension of every agentâ€™s coverage stays unchanged.

With such a minimal parameterization overhead, AR coverage considerably improves the illustration energy of PG strategies. We comment that PG with AR coverage (PG-AR) can concurrently signify all optimum coverage modes within the permutation recreation.

Determine: the heatmaps of actions for insurance policies discovered by PG-Ind (left) and PG-AR (center), and the heatmap for rewards (proper); whereas PG-Ind solely converge to a selected mode within the 4-player permutation recreation, PG-AR efficiently discovers all of the optimum modes.

In additional advanced environments, together with SMAC and GRF, PG-AR can study fascinating emergent behaviors that require sturdy intra-agent coordination that will by no means be discovered by PG-Ind.

Determine 6: (left) emergent conduct induced by PG-AR in SMAC and GRF. On the 2m_vs_1z map of SMAC, the marines hold standing and assault alternately whereas making certain there is just one attacking marine at every timestep; (proper) within the academy_3_vs_1_with_keeper situation of GRF, brokers study a “Tiki-Taka” fashion conduct: every participant retains passing the ball to their teammates.

## Discussions and Takeaways

On this submit, we offer a concrete evaluation of VD and PG strategies in cooperative MARL. First, we reveal the limitation on the expressiveness of fashionable VD strategies, displaying that they may not signify optimum insurance policies even in a easy permutation recreation. In contrast, we present that PG strategies are provably extra expressive. We empirically confirm the expressiveness benefit of PG on fashionable MARL testbeds, together with SMAC, GRF, and Hanabi Problem. We hope the insights from this work may gain advantage the group in the direction of extra basic and extra highly effective cooperative MARL algorithms sooner or later.

*This submit is predicated on our paper: Revisiting Some Frequent Practices in Cooperative Multi-Agent Reinforcement Studying (paper, web site).*