bdss_pomp.c File Reference

#include "pomplink.h"
#include "internal.h"

Include dependency graph for bdss_pomp.c:

Go to the source code of this file.

Macros
#define	Normal   1
#define	Superspreader   2
#define	lambda_nn   (__p[__parindex[0]])
#define	lambda_ns   (__p[__parindex[1]])
#define	lambda_sn   (__p[__parindex[2]])
#define	lambda_ss   (__p[__parindex[3]])
#define	mu   (__p[__parindex[4]])
#define	chi   (__p[__parindex[5]])
#define	POP   (__p[__parindex[6]])
#define	N0   (__p[__parindex[7]])
#define	S0   (__p[__parindex[8]])
#define	N   (__x[__stateindex[0]])
#define	S   (__x[__stateindex[1]])
#define	ll   (__x[__stateindex[2]])
#define	node   (__x[__stateindex[3]])
#define	ell1   (__x[__stateindex[4]])
#define	ell2   (__x[__stateindex[5]])
#define	COLOR   (__x[__stateindex[6]])
#define	EVENT_RATES
#define	lik   (__lik[0])

Functions
static int	random_choice (double n)
static void	change_color (double *color, int nsample, int n, int from, int to)
static double	event_rates (double *__x, const double *__p, double t, const int *__stateindex, const int *__parindex, const int *__covindex, const double *__covars, double *rate, double *logpi, double *penalty)
void	bdss_rinit (double *__x, const double *__p, double t0, const int *__stateindex, const int *__parindex, const int *__covindex, const double *__covars)
void	bdss_gill (double *__x, const double *__p, const int *__stateindex, const int *__parindex, const int *__covindex, const double *__covars, double t, double dt)
void	bdss_dmeas (double *__lik, const double *__y, const double *__x, const double *__p, int give_log, const int *__obsindex, const int *__stateindex, const int *__parindex, const int *__covindex, const double *__covars, double t)
	Measurement model likelihood (dmeasure).

Variables
static const int	nrate = 8

Macro Definition Documentation

chi

#define chi   (__p[__parindex[5]])

Definition at line 31 of file bdss_pomp.c.

COLOR

#define COLOR   (__x[__stateindex[6]])

Definition at line 41 of file bdss_pomp.c.

ell1

#define ell1   (__x[__stateindex[4]])

Definition at line 39 of file bdss_pomp.c.

ell2

#define ell2   (__x[__stateindex[5]])

Definition at line 40 of file bdss_pomp.c.

EVENT_RATES

#define EVENT_RATES

Value:

  event_rates(__x,__p,t,                                \
              __stateindex,__parindex,__covindex,       \
              __covars,rate,logpi,&penalty)             \

Definition at line 43 of file bdss_pomp.c.

#define EVENT_RATES                                     \
  event_rates(__x,__p,t,                                \
              __stateindex,__parindex,__covindex,       \
              __covars,rate,logpi,&penalty)             \
 

lambda_nn

#define lambda_nn   (__p[__parindex[0]])

Definition at line 26 of file bdss_pomp.c.

lambda_ns

#define lambda_ns   (__p[__parindex[1]])

Definition at line 27 of file bdss_pomp.c.

lambda_sn

#define lambda_sn   (__p[__parindex[2]])

Definition at line 28 of file bdss_pomp.c.

lambda_ss

#define lambda_ss   (__p[__parindex[3]])

Definition at line 29 of file bdss_pomp.c.

lik

#define lik   (__lik[0])

Definition at line 389 of file bdss_pomp.c.

ll

#define ll   (__x[__stateindex[2]])

Definition at line 37 of file bdss_pomp.c.

mu

#define mu   (__p[__parindex[4]])

Definition at line 30 of file bdss_pomp.c.

N

#define N   (__x[__stateindex[0]])

Definition at line 35 of file bdss_pomp.c.

N0

#define N0   (__p[__parindex[7]])

Definition at line 33 of file bdss_pomp.c.

node

#define node   (__x[__stateindex[3]])

Definition at line 38 of file bdss_pomp.c.

Normal

#define Normal   1

Definition at line 4 of file bdss_pomp.c.

POP

#define POP   (__p[__parindex[6]])

Definition at line 32 of file bdss_pomp.c.

S

#define S   (__x[__stateindex[1]])

Definition at line 36 of file bdss_pomp.c.

S0

#define S0   (__p[__parindex[8]])

Definition at line 34 of file bdss_pomp.c.

Superspreader

#define Superspreader   2

Definition at line 5 of file bdss_pomp.c.

Function Documentation

bdss_dmeas()

void bdss_dmeas	(	double *	__lik,
		const double *	__y,
		const double *	__x,
		const double *	__p,
		int	give_log,
		const int *	__obsindex,
		const int *	__stateindex,
		const int *	__parindex,
		const int *	__covindex,
		const double *	__covars,
		double	t )

Measurement model likelihood (dmeasure).

Definition at line 392 of file bdss_pomp.c.

   {
  assert(!ISNAN(ll));
  lik = (give_log) ? ll : exp(ll);
}

bdss_gill()

void bdss_gill	(	double *	__x,
		const double *	__p,
		const int *	__stateindex,
		const int *	__parindex,
		const int *	__covindex,
		const double *	__covars,
		double	t,
		double	dt )

Simulator for the latent-state process (rprocess).

This is the Gillespie algorithm applied to the solution of the filter equation for the BDSS process.

Definition at line 162 of file bdss_pomp.c.

  {
  double tstep = 0.0, tmax = t + dt;
  double *color = &COLOR;
  const int nsample = *get_userdata_int("nsample");
  const int *nodetype = get_userdata_int("nodetype");
  const int *lineage = get_userdata_int("lineage");
  const int *sat = get_userdata_int("saturation");
  const int *index = get_userdata_int("index");
  const int *child = get_userdata_int("child");
 
  int parent = (int) nearbyint(node);
 
#ifndef NDEBUG
  int nnode = *get_userdata_int("nnode");
  assert(parent>=0);
  assert(parent<=nnode);
#endif
 
  int parlin = lineage[parent];
  int parcol = color[parlin];
  assert(parlin >= 0 && parlin < nsample);
 
  ll = 0;
 
  // singular portion of filter equation
  switch (nodetype[parent]) {
  default:
    break;
  case 0:                       // root
    assert(sat[parent]==1);
    {
      int c = child[index[parent]];
      assert(lineage[parent]==lineage[c]);
      if (N-ell1 + S-ell2 > 0) {
        double x = (N-ell1)/(N-ell1 + S-ell2);
        if (unif_rand() < x) {
          color[lineage[c]] = Normal;
          ell1 += 1;
          ll -= log(x);
        } else {
          color[lineage[c]] = Superspreader;
          ell2 += 1;
          ll -= log(1-x);
        }
      } else {
        ll += R_NegInf;
        if (unif_rand() < 0.5) {
          color[lineage[c]] = Normal;
          ell1 += 1; N += 1;
        } else {
          color[lineage[c]] = Superspreader;
          ell2 += 1; S += 1;
        }
      }
    }
    break;
  case 1:                       // sample (destructive only)
    if (sat[parent] == 0) {
      if (parcol == Normal) {
        ell1 -= 1;
        ll += (N > 0) ? log(chi*N) : R_NegInf;
        N -= 1;
      } else if (parcol == Superspreader) {
        ell2 -= 1;
        ll += (S > 0) ? log(chi*S) : R_NegInf;
        S -= 1;
      } else {
        ll += R_NegInf;
      }
    } else {
      assert(0);                // #nocov
      ll += R_NegInf;           // #nocov
    }
    color[parlin] = R_NaReal;
    break;
  case 2:                       // branch point
    assert(sat[parent]==2);
    {
      int c1 = child[index[parent]];
      int c2 = child[index[parent]+1];
      assert(c1 != c2);
      assert(lineage[c1] != lineage[c2]);
      assert(lineage[c1] != parlin || lineage[c2] != parlin);
      assert(lineage[c1] == parlin || lineage[c2] == parlin);
 
      if (parcol == Normal) {
        double lam = lambda_nn + lambda_ns;
        if (lam <= 0) { ll += R_NegInf; break; }
        double p_nn = lambda_nn/lam;
        if (unif_rand() < p_nn) {
          // N->N birth: both children Normal
          N += 1; ell1 += 1;
          ll += log(2*lam/N);
          if (N*(N-1) <= 0) ll += R_NegInf;
          color[lineage[c1]] = Normal;
          color[lineage[c2]] = Normal;
        } else {
          // N->S birth: one N, one S
          S += 1; ell2 += 1;
          ll += log(lam/S);
          if (N*S <= 0) ll += R_NegInf;
          if (unif_rand() < 0.5) {
            color[lineage[c1]] = Normal;
            color[lineage[c2]] = Superspreader;
          } else {
            color[lineage[c1]] = Superspreader;
            color[lineage[c2]] = Normal;
          }
          ll -= log(0.5);
        }
      } else if (parcol == Superspreader) {
        double lam = lambda_sn + lambda_ss;
        if (lam <= 0) { ll += R_NegInf; break; }
        double p_ss = lambda_ss/lam;
        if (unif_rand() < p_ss) {
          // S->S birth: both children Superspreader
          S += 1; ell2 += 1;
          ll += log(2*lam/S);
          if (S*(S-1) <= 0) ll += R_NegInf;
          color[lineage[c1]] = Superspreader;
          color[lineage[c2]] = Superspreader;
        } else {
          // S->N birth: one S, one N
          N += 1; ell1 += 1;
          ll += log(lam/N);
          if (N*S <= 0) ll += R_NegInf;
          if (unif_rand() < 0.5) {
            color[lineage[c1]] = Normal;
            color[lineage[c2]] = Superspreader;
          } else {
            color[lineage[c1]] = Superspreader;
            color[lineage[c2]] = Normal;
          }
          ll -= log(0.5);
        }
      } else {
        ll += R_NegInf;
      }
    }
    break;
  }
 
  // regular portion of filter equation
  if (tmax > t && R_FINITE(ll)) {
 
    double rate[nrate], logpi[nrate];
    int event;
    double event_rate = 0;
    double penalty = 0;
 
    event_rate = EVENT_RATES;
    tstep = exp_rand()/event_rate;
 
    while (t + tstep < tmax) {
      event = rcateg(event_rate,rate,nrate);
      assert(event>=0 && event<nrate);
      ll -= penalty*tstep + logpi[event];
      switch (event) {
      case 0:                   // N->N birth
        N += 1;
        assert(!ISNAN(ll));
        break;
      case 1:                   // N->S birth, s=(0,0) or s=(0,1)
        S += 1;
        ll += log(1-ell2/S);
        assert(!ISNAN(ll));
        break;
      case 2:                   // N->S birth, s=(1,0)
        change_color(color,nsample,random_choice(ell1),Normal,Superspreader);
        ell1 -= 1; ell2 += 1;
        S += 1;
        ll += log(1-ell1/N)-log(S);
        assert(!ISNAN(ll));
        break;
      case 3:                   // S->S birth
        S += 1;
        assert(!ISNAN(ll));
        break;
      case 4:                   // S->N birth, s=(0,0) or s=(1,0)
        N += 1;
        ll += log(1-ell1/N);
        assert(!ISNAN(ll));
        break;
      case 5:                   // S->N birth, s=(0,1)
        change_color(color,nsample,random_choice(ell2),Superspreader,Normal);
        ell2 -= 1; ell1 += 1;
        N += 1;
        ll += log(1-ell2/S)-log(N);
        assert(!ISNAN(ll));
        break;
      case 6:                   // death N
        assert(N>=1);
        N -= 1;
        break;
      case 7:                   // death S
        assert(S>=1);
        S -= 1;
        break;
      default:                  // #nocov
        assert(0);              // #nocov
        break;                  // #nocov
      }
 
      ell1 = nearbyint(ell1);
      ell2 = nearbyint(ell2);
 
      t += tstep;
      event_rate = EVENT_RATES;
      tstep = exp_rand()/event_rate;
 
    }
    tstep = tmax - t;
    ll -= penalty*tstep;
  }
  node += 1;
}

Here is the call graph for this function:

bdss_rinit()

void bdss_rinit	(	double *	__x,
		const double *	__p,
		double	t0,
		const int *	__stateindex,
		const int *	__parindex,
		const int *	__covindex,
		const double *	__covars )

Latent-state initializer (rinit component).

The state variables include N, S plus 'ell1' and 'ell2' (numbers of N- and S-deme lineages), the accumulated weight ('ll'), the current node number ('node'), and the coloring of each lineage ('COLOR').

Definition at line 128 of file bdss_pomp.c.

  {
  if (!R_FINITE(lambda_nn) || lambda_nn < 0 ||
      !R_FINITE(lambda_ns) || lambda_ns < 0 ||
      !R_FINITE(lambda_sn) || lambda_sn < 0 ||
      !R_FINITE(lambda_ss) || lambda_ss < 0 ||
      !R_FINITE(mu) || mu < 0 || !R_FINITE(chi) || chi < 0)
    err("BDSS rates must be finite and nonnegative.");
  if (!R_FINITE(POP) || POP < 0 ||
      !R_FINITE(N0) || N0 < 0 || !R_FINITE(S0) || S0 < 0 ||
      !R_FINITE(N0+S0) || N0+S0 <= 0)
    err("BDSS initial values must be finite and nonnegative, with N0+S0 > 0.");
  double adj = POP/(N0+S0);
  if (!R_FINITE(adj)) err("BDSS: pop/(N0+S0) must be finite.");
  N = nearbyint(N0*adj);
  S = nearbyint(S0*adj);
  ell1 = 0;
  ell2 = 0;
  ll = 0;
  node = 0;
}

change_color()

static void change_color ( double * color, int nsample, int n, int from, int to )

static

Definition at line 13 of file bdss_pomp.c.

                                                   {
  int i = -1;
  while (n >= 0 && i < nsample) {
    i++;
    if (!ISNA(color[i]) && nearbyint(color[i]) == from) n--;
  }
  assert(i < nsample);
  assert(n == -1);
  assert(nearbyint(color[i]) == from);
  color[i] = to;
}

Here is the call graph for this function:

Here is the caller graph for this function:

event_rates()

static double event_rates ( double * __x, const double * __p, double t, const int * __stateindex, const int * __parindex, const int * __covindex, const double * __covars, double * rate, double * logpi, double * penalty )

static

Definition at line 48 of file bdss_pomp.c.

   {
  double event_rate = 0;
  double alpha, pi, disc;
  *penalty = 0;
  assert(N >= ell1 && ell1 >= 0);
  assert(S >= ell2 && ell2 >= 0);
  // 0: N->N birth, s=(0,0) or s=(1,0)
  alpha = lambda_nn*N;
  disc = (N > 0) ? ell1*(ell1-1)/N/(N+1) : 1;
  *penalty += alpha*disc;
  event_rate += (*rate = alpha*(1-disc)); rate++;
  *logpi = 0; logpi++;
  // 1: N->S birth, s=(0,0) or s=(0,1)
  alpha = lambda_ns*N;
  pi = 1/(ell1+1);
  event_rate += (*rate = alpha*pi); rate++;
  *logpi = log(pi); logpi++;
  // 2: N->S birth, s=(1,0)
  pi = ell1/(ell1+1);
  event_rate += (*rate = alpha*pi); rate++;
  *logpi = (ell1 > 0) ? -log(ell1+1) : 0; logpi++;
  // 3: S->S birth, s=(0,0) or s=(0,1)
  alpha = lambda_ss*S;
  disc = (S > 0) ? ell2*(ell2-1)/S/(S+1) : 1;
  *penalty += alpha*disc;
  event_rate += (*rate = alpha*(1-disc)); rate++;
  *logpi = 0; logpi++;
  // 4: S->N birth, s=(0,0) or s=(1,0)
  alpha = lambda_sn*S;
  pi = 1/(ell2+1);
  event_rate += (*rate = alpha*pi); rate++;
  *logpi = log(pi); logpi++;
  // 5: S->N birth, s=(0,1)
  pi = ell2/(ell2+1);
  event_rate += (*rate = alpha*pi); rate++;
  *logpi = (ell2 > 0) ? -log(ell2+1) : 0; logpi++;
  // 6: death N
  alpha = mu*N;
  if (N > ell1) {
    event_rate += (*rate = alpha); rate++;
    *logpi = 0; logpi++;
  } else {
    *rate = 0; rate++;
    *logpi = 0; logpi++;
    *penalty += alpha;
  }
  // 7: death S
  alpha = mu*S;
  if (S > ell2) {
    event_rate += (*rate = alpha); rate++;
    *logpi = 0; logpi++;
  } else {
    *rate = 0; rate++;
    *logpi = 0; logpi++;
    *penalty += alpha;
  }
  // sampling (Q = 0): destructive only
  *penalty += chi*(N+S);
  assert(R_FINITE(event_rate));
  return event_rate;
}

random_choice()

static int random_choice ( double n )

inlinestatic

Definition at line 9 of file bdss_pomp.c.

                                           {
  return floor(R_unif_index(n));
}

Here is the caller graph for this function:

Variable Documentation

nrate

const int nrate = 8

static

Definition at line 7 of file bdss_pomp.c.