Efficient task offloading in UAV-assisted heterogeneous mobile edge computing (MEC) networks is increasingly challenged by the co-existence of operationally distinct workload scenarios—including high-demand bursts, resource-constrained periods, and balanced operational states—each demanding fundamentally different assignment strategies. In such networks, mobile executor nodes (e.g., UAVs or vehicle-mounted edge servers) must be dispatched to the vicinity of geographically distributed tasks, making assignment decisions jointly dependent on node mobility, the quality of sensing data, and dynamic resource availability. Conventional approaches based on combinatorial optimization with fixed parameters or greedy heuristics fail to adapt to these varying conditions, leading to resource depletion under sequential workloads or underutilization under high-demand bursts. To address these limitations, this paper proposes SAGE (Scene-Adaptive Graph-Enhanced offloading), a task-offloading framework that combines a heterogeneous graph neural network (HeteroGNN) with a dueling double DQN meta-controller and a mixed-integer linear programming (MILP) solver. At the state-representation level, a heterogeneous bipartite graph is constructed over mobile executor nodes and tasks, with type-specific projection layers encoding the semantic features of each node type and three-dimensional edge features—comprising task success probability, normalized service distance, and link quality—integrated via edge-gated message passing. At the decision level, the meta-controller perceives the current workload scenario through a seven-dimensional situational state vector fused with the graph embedding, selects an appropriate offloading strategy from a learned discrete action space, and drives the MILP solver to perform task-chain assignment under the selected configuration. Experiments on 60 fixed evaluation episodes spanning three representative workload scenarios demonstrate that SAGE achieves an overall reward improvement of 15.9% over the best fixed-strategy baseline, reduces the resource depletion rate to 16.7%, and maintains a high-priority task completion rate of 84.7%. Particularly under resource-constrained conditions, SAGE reduces the reward deficit by 72.3% relative to the best fixed strategy (from to ), demonstrating strong scene-adaptive decision-making capability.
Classical Colonel Blotto models predict smooth and monotone allocation patterns under a homogeneous-resource benchmark. This paper studies how heterogeneous technologies and partially irreversible deployment reshape equilibrium structure through a unifying reduced-form mechanism, which we call the induced cost geometry. In the compatible linear case developed in the main text, heterogeneous technologies and irreversibility generate battlefield-specific effective marginal costs , where is the endogenous shadow price of the global budget. We distinguish results that rely on this compatible linear representation from those that extend directly to the more general lower-envelope induced cost object. The induced cost object serves as the central organizing device for the equilibrium analysis and yields three main structural implications. First, under binding budgets, equilibrium effort can be sparse: positive effort is concentrated on battlefields whose effective marginal costs remain sufficiently low relative to their values, generating selective participation even under smooth contest success functions. Second, within a class of local fixed-support equilibrium configurations with binding budgets, an increase in a battlefield’s value can reduce equilibrium effort on that battlefield via the shadow-price channel—a mechanism-identification result for a specific structural regime rather than a general comparative-statics claim. Third, along durability paths, contestability and equilibrium support can exhibit discrete regime changes generated by technology switching within a finite lower-envelope cost structure. We develop a unified reduced-form characterization of equilibrium structure via induced costs and KKT conditions, highlighting how induced cost geometry reshapes both the intensive and extensive margins of strategic allocation within the equilibrium regimes characterized in the paper.
The Colonel Blotto game is a classical model of competitive resource allocation, but equilibrium computation becomes difficult in heterogeneous and asymmetric instances. In this research, we study a finite-horizon dynamic reformulation in which players allocate resources sequentially over publicly observed stages, while the payoff depends only on terminal cumulative allocations. The purpose of the reformulation is not to change the primitive objective, but to represent the same terminal-payoff problem as a zero-sum Markov game. We first show that the dynamic formulation admits a pathwise payoff-equivalent Markov representation through telescoping rewards. Under a known finite horizon, costless carryover, and terminal-only payoff evaluation, the dynamic game and the corresponding static Blotto game have the same minimax value at every reachable continuation state. This is a value-equivalence result; it does not imply a one-to-one correspondence between static and dynamic equilibrium strategy sets. The proof is based on terminal-deferral upper and lower bounds for the two players. We also study action-independent geometric termination, for which the discounted telescoping return coincides exactly with the expected stopped terminal payoff, and we provide a probability-controlled mismatch bound for truncated stopping rules. Numerical finite-grid experiments illustrate the value identity and report residual diagnostics. The results clarify when sequential Markov representations preserve the original Blotto objective and when additional primitives, such as carryover depreciation or primitive flow payoffs, require separate analysis.
Translating natural language into graph query languages (NL2GQL) enables non-expert users to access graph databases, but supervised parsers depend on large annotated corpora and costly retraining whenever the schema evolves. Few-shot in-context learning offers a training-free alternative, yet its effectiveness hinges on demonstration selection, and conventional retrieval based on sentence-level matching and a single similarity signal often overlooks beneficial exemplars. We propose MGDR-SDR, a training-free approach that applies constituency parsing to derive three complementary representations of the utterance: a shallow abstraction preserving fine-grained lexical and syntactic detail, a deep abstraction exposing the intent-level syntactic skeleton, and a noun phrase list capturing concrete graph elements. These representations drive a multi-granular retrieval procedure that aggregates evidence across dense and sparse channels, yielding broader and more precise demonstration coverage. We evaluate MGDR-SDR on SPARQL generation over KQA Pro and Cypher generation over ZOGRASCOPE. Without any training, it reaches 86.58% Hits@1 on KQA Pro, surpassing strong supervised parsers and setting a new state-of-the-art among few-shot methods, and it transfers across GQL formalisms, attaining the best iid accuracy and a balanced generalization profile that surpasses all open-source few-shot baselines on ZOGRASCOPE. These results show that syntax-derived, multi-granular retrieval is an effective and generalizable strategy for few-shot NL2GQL semantic parsing.