Traveling Salesman Problem, multiple features with ragged tensor

TensorFlow General Discussion
,
1

Hello, I am trying to solve traveling salesman problem TSP, finding the shortest route between locations [location id, coord x, coord y]

locations=[
['a',0,0],
['b',3,0],
['c',0,4]
]

But there could be more than 3 locations to be inputted so I found tensor.ragged
This is for training 2 and 3 locations TSP,

inputs=tf.ragged.constant([[[0,0],[3,0]],#a,b
[[0,0],[3,0],[0,4]]])#a,b,c
outputs= tf.ragged.constant([[[0,0],[3,0]],#a,b
[[0,4],[3,0],[0,0]]])#c,b,a
model = tf.keras.Sequential([
    tf.keras.layers.Input(shape=[None], dtype=tf.int64, ragged=True),
    tf.keras.layers.Dense(50,activation='relu')
])
model.compile(loss='mean_absolute_error', optimizer='rmsprop')
model.fit(inputs, outputs, epochs=5)
print(model.predict(inputs))

But I got this error, how to tell dense layer it is ragged tensor? Or specifically, how to build the layer? Thank you

    raise ValueError('The last dimension of the inputs to a Dense layer '
ValueError: The last dimension of the inputs to a Dense layer should be defined. Found None. Full input shape received: (None, None)
3

Hello @ZacharyLaw

Thank you for using TensorFlow
Ragged tensors are not entirely supported by Dense layers, if you have a Dataset that contains non-ragged tensors, and tensor lengths vary across elements, then you can batch those non-ragged tensors into ragged tensors by applying the dense_to_ragged_batch transformation.
In your case best way to deal with this Traveling salesman problem is to find suitable RNN like LSTM and proceed with the development of solution.