Model convergence problem; function evaluation limit reached without convergence

[loulx]

Hello,
I am trying to run a glmmTMB model with a binomial variable as a response variable (0 or 1), and my variable intersection that is also binomial (0 or 1) and I get the following warning:
Model convergence problem; function evaluation limit reached without convergence (9).
Here is my code:

TMBStruc = glmmTMB(case_ ~ -1 + habitat1*intersection + habitat2*intersection+ habitat3*intersection
                   +habitat3*intersection + infrastructures*intersection2 + time*intersection2 + season*intersection2 +
                      (1|id) +
                     (0 + annee | id),
                   family=binomial, data=tabgTMB, doFit=FALSE)
TMBStruc1 = update(TMBStruc, control=glmmTMBControl(optimizer=optim,
                                                    optArgs=list(method="BFGS")))

TMBStruc1$parameters$theta[1] = log(1e3)

nvarparm <- length(TMBStruc1$parameters$theta)
TMBStruc1$mapArg = list(theta=factor(c(NA,1:(nvarparm-1))))

glmm.TMB.random <- glmmTMB:::fitTMB(TMBStruc1)
summary(glmm.TMB.random)

modele1 <- summary(glmm.TMB.random)

I tried to add an optimizer but it does not work, would you have any idea please?

[mwild]

Hi ,
I have a similar problem with convergence but did not manage to resolve it.
"Model convergence problem; non-positive-definite Hessian matrix."