Abstract: The present disclosure relates to an afterburner structure with fuel injection nozzles. In the condition of a stable inlet flow, the present disclosure forms a sweeping oscillating jet with a certain frequency at the outlet under an alternate feedback action of a feedback channel and the fluid's Coanda effect, so that the spatial uniformity of the fuel injection is improved; Therefore, without increasing the structural complexity of current afterburner of the turbo-engine and the combustion chamber of the subsonic combustion ramjet, the atomization performance and the uniformity of the oil-gas mixture in the afterburner/combustion chamber of the ramjet can be greatly improved, the combustion efficiency and combustion instability of the engine can be improved and meanwhile the length of the afterburner and the combustion chamber of the subsonic combustion ramjet is shortened.

Abstract: According to examples, an apparatus may include a processor and a memory on which is stored machine readable instructions that may cause the processor to generate a matrix based on features of reference entities and a respective allocation of value provided to each reference entity, apply first-stage machine-learning on the matrix to identify relevant features of the reference entities that correlate with the respective allocation of value provided to each reference entity, and access an identity of a target entity and target features of the target entity. The instructions may further cause the processor to apply second-stage machine-learning to generate a cluster comprising the target entity and a set of the reference entities based on the relevant features and determine a distribution of values allocated to the set. The distribution of values may be used to generate or assess a target allocation of value for the target entity.

Abstract: According to examples, an apparatus may include a processor and a memory on which is stored machine readable instructions that may cause the processor to generate a matrix based on features of reference entities and a respective allocation of value provided to each reference entity, apply first-stage machine-learning on the matrix to identify relevant features of the reference entities that correlate with the respective allocation of value provided to each reference entity, and access an identity of a target entity and target features of the target entity. The instructions may further cause the processor to apply second-stage machine-learning to generate a cluster comprising the target entity and a set of the reference entities based on the relevant features and determine a distribution of values allocated to the set. The distribution of values may be used to generate or assess a target allocation of value for the target entity.