During the metabolism (sugars, proteins, fats) in tissue the heat is generated...

Nikola thanks, but what is the mechanism by which the body temperature remains almost constant at 37 oC without much fluctuation.  

Via centers located in preoptical area and frontal nuclei of hypothalamus. Receptors, synapses and neurons...I dont understand what is so unclear. Like hormones and other signal molecules tiger different systems (depending on temperature) causing in one case vasodilatation, sweating and decrease in heat generation or on the other hand vasoconstrition, pilo erection and termogenesis through metabolism.

What can be unclear? Secondary messengers that cause all those reactions? or what?

I'm wondering if what you're driving at is how homeotherms maintain relatively constant heat generation at a molecular level? The reason why mammals and birds generate metabolic heat but non-avian reptiles, fish, amphibians do not (or not as much, they still produce a little), is that endotherms have leaky membranes in their cells. In particular endotherms leak Sodium and Potassium the wrong way across the membrane and have to do constant work to pull the Sodium and Potassium back to where they should be to maintain a gradient. The constant pumping of Na and K generates heat as a byproduct. It's interesting from an evolutionary point of view, because what would probably have been initially a mutation that resulted in a huge loss of efficiency (lizards are way more metabolically efficient than we are) generated a heat byproduct, that in turn allowed for greater activity, that in turn allowed for eating more food to fuel the inefficient metabolism. It's sort like having taken a step backwards in efficiency and discovering that it still works out anyway, especially in colder climates where the heat byproduct more than compensates for the need to eat more.

Probably the most insightful paper in this area is that from Prof Nakamura:

http://www.ncbi.nlm.nih.gov/pubmed/21900642

It is also open access. 

To summarize this review, there are three main "parts" which contribute to effective thermoregulation in mammals:

1) Sensory

2) Integration 

3) Effector 

1) Sensation

It is generally accepted now that TRP ion channels are responsible for thermosensory processes. The TRP superfamily includes a range of proteins which become active in response to different environmental temperatures. For example, TRPV1 is activated in response to noxious heat, TRPM8 becomes activated in response to environmental cold, etc etc. The mechanisms for their activation are not entirely accepted by may be due to temperature mediated structural modifications of the protein. 

These TRP channels are located not only in the epidermis, but are also widely distributed throughout splanchnic and vagus nerve afferent fibres distributed in the abdomen. It is interesting that there is ~ a 9 times more effective thermoregulatory response to a 1 degree Celsius change in core temperature compared with that of skin temperature. I am not 100% sure why this is, but my educated guess would be that TPR channels are simply more condensed in this region (and thus there is a stronger signal 'delivered' to the hypothalamus). 

Briefly, activation of either cooling or warming signals sensed by TRP channels activate somatosensory neurons that innervate neurons in the spinal and trigeminal dorsal horns. 

The below book chapter is a nice introduction to the above:

http://www.ncbi.nlm.nih.gov/books/NBK5244/

2) Integration of the thermosensory signal

Contrary to some belief, the hypothalamus is not the only structure in the brain involved in effective thermoregulation. Although signals relayed to the hypothalamus are vitally important for involuntary, autonomic responses, the thermosensory signals are also relayed to the somatosensory cortex.  This allows for concious discrimination of the environment (external of internal) and allows for 'behavioural thermoregulation'. However, I won't expand any more on that here. 

In order for autonomic responses to be activated, thermosensory signals must be relayed to the pre-optic area. There have been many lesion and retrograde neural tracing studies  which have confirmed the pre-optic area to be essential for thermoregulatory responses to hot and cold. Warm neurons that project to the pre-optic area originate in the lateral parabrachial nucleus, dorsal subregion and cool neurons project to the pre-optic area from the lateral parabrachial nucleus, external lateral subregion. The neurotransmitter glutamate mediates autonomic thermogenic and heat loss responses (blockade of glutamate receptors inhibits both responses and nanoinjection of NMDA in the lateral parabrachial nucleus activates both responses). 

There are other relay nuclei within the pre-optic area, such as the median pre-optic nucleus and the medial pre-optic area, but for the sake of this answer I won't expand. So far we have a signalling pathway consisting of:

Activation of thermosensory neuron

Synapse to dorsal horn 

Synapse to lateral parabrachial nucleus

Signal relayed to pre-optic area (median pre-optic nucleus and and medial pre-optic area) 

3) Effector responses

GABAergic projections from the medial pre-optic area tonically inhibit the activities of neurons in sympathoexcitatory sites. Again, there a several populations of neurons which are inhibited under these conditions (dorsomedial hypothalamus, rostral medullary raphe). If these neuronal population are dis-inhibited (this can be achieved by micro-injection of the GABA receptor agonist bicuculline), then thermoegenic responses such as shivering, tachycardia and cutaneous vasoconstriction arise. It is interesting that during fever, activation of the EP3 receptor on these GABAergic neurons inhibits cyclic AMP and causes these neurons to become dis-inhibited.

Descending motor outputs travel from the spinal chord (via the ventral horn) and synapse with neurons connected to skeletal muscle, brown adipose tissue, and the cutaneous circulation. Acetylcholine mediates shivering responses and also plays a role in skin vasoconstriction. Norepinephrine is largely responsible for brown adipose tissue thermogenesis, but the role for BAT thermogenesis in adult humans is still up for debate.

Hope that helps in some way!

Another approach:

Computer buffs know that artificial neural nets employ multi-layer nets which utilize the technique of back propagation of errors.  Backpropagation approximates the non-linear relationship between the input and the output by adjusting the weighting values internally. The brain is a massively parallel processor and possibly uses such a technique (in the hindbrain) for the phenomena of ‘homeostasis’ (regulation of body temperature, respiration, etc.) where a fixed output is desired irrespective of the variability of the inputs.