Wator e' un semplice modello di andamento della popolazione acquatica studiato in ecologia. Questo modello e' diventato popolare dopo esser stato presentato nella colonna Computer Recreations della rivista Scientific American [A. K. Dewdney, Sharks and fish wage an ecological war on the toroidal planet Wator, Scientific American, December 1984]. Wator e' un pianeta ricoperto interamente da un oceano temperato. L'oceano e' abitato da predatori (gli squali)e prede (i pesci). Wator viene rappresentato da una matrice rettangolare. Ogni elemento della matrice rappresenta un'area dell'oceano che puo' essere in tre stati distinti

• puo' essere vuota
• puo' contenere esattamente un pesce
• puo' contenere esattamente uno squalo.

Il tempo e' a sua volta suddiviso in unita' standard dette chronons, il sistema si evolve in passi (step) che corrispondono ad un chronon ciascuno. Ad ogni passo la matrice che rappresenta il pianeta viene aggiornata secondo il comportamento di predatori e prede. Il comportamento di predatori e prede e' il seguente:

• Gli squali mangiano e si spostano. A ogni passo, uno squalo presente in una cella (i,j) si guarda intorno nelle celle adiacenti. Se una di queste celle contiene un pesce, lo squalo mangia il pesce e si sposta nella cella precedentemente occupata dal pesce. Se nessuna delle celle adiacenti contiene un pesce, lo squalo si sposta in una delle celle adiacenti vuote. Le celle adiacenti sono le celle in cui uno degli indici differisce esattamente di uno da i e j, graficamente

		*
	*		*
		*

• Sharks eat and move. At each step, a shark will look around in all the adjacent (nearest neighbor ) cells, and eat a fish if it finds one there; then the shark will move to the space previously occupied by the fish. If the shark cannot find any fish, it will move into one of the empty adjacent cells, if there any. [In a rectangular matrix, the nearest-neighbor elements are those whose first or second index (but not both) differ by exactly one.

• Sharks breed and die. If a shark survives for Sb or more chronons, it will breed: if an empty adjacent cell is available, a new shark will appear there. The parent will breed again after Sb chronons. However, if a shark has not eaten after Sd chronons, it will die (disappear from the grid).

• Fish move. At each step, each fish will move at random into one of the free adjacent cells, if there are any. Since Wator is toroidal, fish (and sharks) can swim out of one edge of the matrix, and come back from the opposite edge

• Fish breed. If a fish survives for Fb chronons or more, it will breed: if an empty adjacent cell is available, a new fish will appear there. The parent will breed again after Fb chronons

• Initial population. At the beginning of the simulation, the grid should be filled up randomly with NS sharks and NF fish. Their breeding and starving ages can all be set to zero

• Random movement. Whenever a shark or fish can choose to operate (move, eat, spawn) in more than one direction, it should do so randomly

These rules are sufficient to create a simple model of predator-prey interaction that displays a behavior similar to that observed in some populations in nature. . When there are enough fish to go around, the sharks can prosper and breed: but if the number of sharks grows too large, they will eat most of the fish and a famine will ensue.

In this assignment, you are going to write a parallel program that implements the Wator model by subdividing (dynamically, for load-balancing purposes) the rectangular matrix that represents the ocean between the parallel processors.